| *** changbl has quit IRC | 00:02 | |
| *** dfarrell07 has joined #openstack-dev | 00:02 | |
| *** morazi has joined #openstack-dev | 00:03 | |
| *** amcrn has quit IRC | 00:04 | |
| *** harlowja is now known as harlowja_away | 00:04 | |
| *** kgriffs is now known as kgriffs_afk | 00:06 | |
| *** carl_baldwin has quit IRC | 00:07 | |
| *** venkatesh has quit IRC | 00:11 | |
| *** amcrn has joined #openstack-dev | 00:11 | |
| *** mriedem has quit IRC | 00:11 | |
| *** eglynn has joined #openstack-dev | 00:11 | |
| *** rodrigods has quit IRC | 00:13 | |
| *** byeager has quit IRC | 00:13 | |
| *** harlowja_away is now known as harlowja | 00:15 | |
| *** RajeshMohan has quit IRC | 00:17 | |
| *** pmathews has quit IRC | 00:17 | |
| *** mrodden has quit IRC | 00:18 | |
| *** matsuhashi has joined #openstack-dev | 00:19 | |
| *** csaba|afk is now known as csaba | 00:20 | |
| *** eglynn has quit IRC | 00:20 | |
| *** sweston has joined #openstack-dev | 00:21 | |
| *** enikanorov_ has joined #openstack-dev | 00:21 | |
| *** giulivo has quit IRC | 00:21 | |
| *** asalkeld has quit IRC | 00:22 | |
| *** anniec has joined #openstack-dev | 00:22 | |
| *** MaxV has quit IRC | 00:22 | |
| *** enikanorov has quit IRC | 00:23 | |
| *** atiwari has quit IRC | 00:23 | |
| *** tjones has quit IRC | 00:23 | |
| *** mrda is now known as mrda_away | 00:23 | |
| *** RajeshMohan has joined #openstack-dev | 00:23 | |
| *** tjones has joined #openstack-dev | 00:23 | |
| *** zyluo has joined #openstack-dev | 00:24 | |
| zyluo | bnemec, ping | 00:24 |
|---|---|---|
| *** FunnyLookinHat has quit IRC | 00:25 | |
| *** flaper87 is now known as flaper87|afk | 00:27 | |
| *** sweston has quit IRC | 00:27 | |
| *** tjones has quit IRC | 00:28 | |
| *** CaptTofu has quit IRC | 00:29 | |
| *** CaptTofu has joined #openstack-dev | 00:30 | |
| *** gokrokve has quit IRC | 00:30 | |
| *** marcoemorais has quit IRC | 00:31 | |
| *** marcoemorais has joined #openstack-dev | 00:31 | |
| *** sweston has joined #openstack-dev | 00:32 | |
| *** CaptTofu has quit IRC | 00:34 | |
| *** dfarrell07 has quit IRC | 00:38 | |
| *** jckasper has joined #openstack-dev | 00:39 | |
| *** IanGovett has quit IRC | 00:40 | |
| *** IanGovett has joined #openstack-dev | 00:40 | |
| *** cadenzajon has quit IRC | 00:41 | |
| *** sarob has joined #openstack-dev | 00:41 | |
| *** CaptTofu has joined #openstack-dev | 00:41 | |
| *** jf-jenni has joined #openstack-dev | 00:42 | |
| *** thuc has quit IRC | 00:42 | |
| *** thuc has joined #openstack-dev | 00:43 | |
| *** zzelle_ has quit IRC | 00:44 | |
| *** thuc_ has joined #openstack-dev | 00:45 | |
| *** sarob has quit IRC | 00:46 | |
| *** spzala has quit IRC | 00:46 | |
| *** thuc has quit IRC | 00:47 | |
| *** lcheng_ has joined #openstack-dev | 00:47 | |
| *** devoid has quit IRC | 00:50 | |
| *** sarob has joined #openstack-dev | 00:51 | |
| *** dfarrell07 has joined #openstack-dev | 00:52 | |
| *** stevemar has quit IRC | 00:55 | |
| *** sarob has quit IRC | 00:56 | |
| *** kgriffs_afk is now known as kgriffs | 00:58 | |
| *** yamahata has joined #openstack-dev | 00:59 | |
| *** markwash_ has joined #openstack-dev | 00:59 | |
| *** browne has quit IRC | 01:01 | |
| *** pablosan has quit IRC | 01:01 | |
| *** lcheng_ has quit IRC | 01:02 | |
| *** markwash has quit IRC | 01:02 | |
| *** markwash_ is now known as markwash | 01:02 | |
| *** sarob has joined #openstack-dev | 01:02 | |
| *** dfarrell07 has quit IRC | 01:03 | |
| *** nelsnelson has quit IRC | 01:03 | |
| *** zyluo has quit IRC | 01:04 | |
| *** anniec has quit IRC | 01:04 | |
| *** zyluo has joined #openstack-dev | 01:04 | |
| *** pablosan has joined #openstack-dev | 01:05 | |
| *** mkollaro has quit IRC | 01:05 | |
| *** BLZbubba has quit IRC | 01:07 | |
| *** godara has quit IRC | 01:07 | |
| *** BLZbubba has joined #openstack-dev | 01:07 | |
| *** kgriffs is now known as kgriffs_afk | 01:07 | |
| *** sarob has quit IRC | 01:08 | |
| *** tongli has quit IRC | 01:09 | |
| *** ijw has quit IRC | 01:09 | |
| *** mrodden has joined #openstack-dev | 01:09 | |
| *** csaba is now known as csaba|afk | 01:10 | |
| *** anniec has joined #openstack-dev | 01:10 | |
| *** cdub has joined #openstack-dev | 01:10 | |
| jamielennox | dolphm, was there another client release? | 01:11 |
| *** mrodden1 has joined #openstack-dev | 01:12 | |
| *** epim has quit IRC | 01:12 | |
| *** melwitt has quit IRC | 01:12 | |
| *** mrodden has quit IRC | 01:14 | |
| *** sandywalsh has quit IRC | 01:14 | |
| *** enikanorov_ has quit IRC | 01:17 | |
| *** enikanorov has joined #openstack-dev | 01:17 | |
| *** thuc_ has quit IRC | 01:19 | |
| *** thuc has joined #openstack-dev | 01:20 | |
| *** xarses has quit IRC | 01:21 | |
| jamielennox | ayoung-afk: have you changed anything about v2 revocations? | 01:23 |
| *** ayoung-afk is now known as ayoung | 01:23 | |
| ayoung | nope | 01:23 |
| *** xmltok has quit IRC | 01:24 | |
| *** thuc has quit IRC | 01:24 | |
| *** tsekiyam_ has joined #openstack-dev | 01:27 | |
| *** nosnos has joined #openstack-dev | 01:28 | |
| *** smurugesan has quit IRC | 01:29 | |
| ayoung | morganfainberg, jamielennox dstanek_afk can we put this one to bed? https://review.openstack.org/#/c/68548/ | 01:30 |
| *** tsekiyama has quit IRC | 01:30 | |
| *** cnesa has quit IRC | 01:30 | |
| *** novas0x2a|laptop has quit IRC | 01:30 | |
| *** hemna has quit IRC | 01:31 | |
| jamielennox | ayoung: my only question about that one is does it make sense to advertise that over ampq or is it really only relevant internal to keystone | 01:31 |
| ayoung | AAAAAH! | 01:31 |
| ayoung | no idea | 01:31 |
| ayoung | does it matter? | 01:31 |
| jamielennox | ayoung: no idea :) | 01:31 |
| *** tsekiyam_ has quit IRC | 01:31 | |
| morganfainberg | and jamielennox from left field | 01:32 |
| jamielennox | :) | 01:32 |
| morganfainberg | ayoung, i think a disable event on amqp is fine | 01:32 |
| morganfainberg | and likely is more relevant than update | 01:32 |
| jamielennox | i can see that other services would want to know about a disable | 01:32 |
| ayoung | I think that some people will want it | 01:32 |
| morganfainberg | someone might care | 01:32 |
| ayoung | disable a user and shutdown their vms? | 01:32 |
| ayoung | does that have anything to do with this patch? Do we indicate an events availability? | 01:33 |
| *** mrda_away is now known as mrda | 01:33 | |
| *** sweston has quit IRC | 01:33 | |
| ayoung | Feeping Creaturism | 01:33 |
| *** peoplemerge has quit IRC | 01:33 | |
| morganfainberg | eh, | 01:34 |
| morganfainberg | *shrug* | 01:34 |
| ayoung | morganfainberg, this is all your fault | 01:34 |
| *** cdub has quit IRC | 01:34 | |
| jamielennox | i think the whole what goes onto the bus is new to all of us | 01:34 |
| morganfainberg | ayoung, lol | 01:34 |
| ayoung | "you should use notifications" you said "It will be easy" you said | 01:34 |
| morganfainberg | and you're the one that listened! | 01:34 |
| ayoung | Fool me once, shame on you, fool me twice...you can't get fooled again. | 01:35 |
| jamielennox | fool me once, shame on you GOTO 10 | 01:36 |
| ayoung | jamielennox, have you ever programmed in Basic? | 01:36 |
| ayoung | I'm guessing yes | 01:36 |
| jamielennox | eh, not really just the best way to write something like that | 01:37 |
| *** sweston has joined #openstack-dev | 01:37 | |
| ayoung | I was actually looking into Logo as a first language to teach my Son. We've been playing a "board game" called Robot Turtles that is like turtle graphics programming (each card is an instruction to move the turtle) and Logo ssems the logical next step | 01:37 |
| ayoung | turns out there is UCBLogo on Fedora with turtle graphics | 01:38 |
| jamielennox | logo was the first thing i got taught | 01:38 |
| ayoung | really? | 01:38 |
| *** thuc has joined #openstack-dev | 01:38 | |
| jamielennox | early high school course | 01:39 |
| jamielennox | country schools thought that it counted as teaching programming | 01:39 |
| *** dims has quit IRC | 01:39 | |
| *** sarob has joined #openstack-dev | 01:39 | |
| *** zyluo has quit IRC | 01:39 | |
| achampion | how old is you son? | 01:39 |
| *** sweston_ has joined #openstack-dev | 01:40 | |
| *** dims has joined #openstack-dev | 01:40 | |
| ayoung | achampion, I'm old, he's young, | 01:40 |
| ayoung | But I am Young | 01:41 |
| *** mst89 has quit IRC | 01:41 | |
| achampion | you = your | 01:41 |
| ayoung | sorry, that should read | 01:41 |
| ayoung | I AM Young | 01:41 |
| *** tqtran has quit IRC | 01:41 | |
| ayoung | achampion, heh | 01:42 |
| *** igor__ has joined #openstack-dev | 01:42 | |
| ayoung | He's 7 | 01:42 |
| achampion | I've been teaching programming to a nephew (10) using robomind | 01:42 |
| *** igor has quit IRC | 01:42 | |
| ayoung | achampion, does it take Logo? | 01:43 |
| *** alop has quit IRC | 01:43 | |
| *** sweston has quit IRC | 01:44 | |
| ayoung | achampion, how do you like it? | 01:44 |
| achampion | no, it's a simple pseudo-code style with basic flow control structures | 01:44 |
| *** bdpayne has quit IRC | 01:44 | |
| *** nati_ueno has quit IRC | 01:44 | |
| *** nati_ueno has joined #openstack-dev | 01:45 | |
| achampion | been reasonable easy to pick up, from simple instructions left, right, and paintWhite, to flow control structure if, repeat and procedures | 01:46 |
| achampion | we did a specific maze solver, then progressed to a general maze solver | 01:47 |
| ayoung | Nice | 01:47 |
| *** nati_uen_ has joined #openstack-dev | 01:48 | |
| ayoung | morganfainberg, you OK with my responses on "Notifications upon disable" | 01:48 |
| *** sarob_ has joined #openstack-dev | 01:48 | |
| achampion | it displays a robot in a universe, with the program visually affecting the robot universe. | 01:48 |
| *** bknudson has joined #openstack-dev | 01:49 | |
| achampion | Next is a copier, then a copier with scale (1/2, double) | 01:49 |
| *** sarob has quit IRC | 01:49 | |
| achampion | http://www.robomind.net/ | 01:49 |
| *** llu has joined #openstack-dev | 01:51 | |
| *** otherwiseguy has joined #openstack-dev | 01:51 | |
| *** nati_ueno has quit IRC | 01:51 | |
| *** ijw has joined #openstack-dev | 01:53 | |
| *** _cjones_ has quit IRC | 01:54 | |
| *** thuc_ has joined #openstack-dev | 01:54 | |
| *** morganfainberg is now known as morganfainberg_Z | 01:54 | |
| *** thuc has quit IRC | 01:58 | |
| *** thuc_ has quit IRC | 01:58 | |
| *** kgriffs_afk is now known as kgriffs | 01:58 | |
| *** newell has quit IRC | 02:02 | |
| ayoung | jamielennox, what is the path to being able to use client certs and Kerberos with the Keystone client? I'm assuming we get auth plugins merged, and we need a plugin for each of those. | 02:02 |
| jamielennox | ayoung: not as much as you'd think - both of those will fall within the transport base session object | 02:03 |
| jamielennox | client certs should be supported already | 02:03 |
| jamielennox | kerberos is not actually all that hard to add | 02:03 |
| ayoung | and we need to shortcircuit the logic to try and pass through the password, and the format for the token request | 02:03 |
| jamielennox | ayoung: yea - so what we would essentially need is a no-op plugin | 02:03 |
| ayoung | no methods="password" | 02:03 |
| jamielennox | auth plugins will fill things into the X-Auth-Token, you just need to get a token | 02:04 |
| jamielennox | if you have no password or whatever just kerberos then you would still have to have something that got a token - you would just not have the plugin submit a password or whatever | 02:05 |
| jamielennox | ayoung: did that make sensse? | 02:05 |
| *** alexpilotti has quit IRC | 02:05 | |
| *** mriedem has joined #openstack-dev | 02:06 | |
| ayoung | jamielennox, what writes the body of the request now? | 02:06 |
| *** ijw has quit IRC | 02:06 | |
| ayoung | passwword plugin? | 02:06 |
| jamielennox | which request? | 02:06 |
| jamielennox | the auth request ? | 02:06 |
| *** dstanek_afk is now known as dstanek | 02:07 | |
| jamielennox | Plugins are essentially just v2 or v3 keystone | 02:07 |
| ayoung | gyee, can you just +2 https://review.openstack.org/#/c/61247/9 so we can move this, and we'll work up the documentation as future work? | 02:07 |
| ayoung | yeah, the body for POST /v3/auth | 02:08 |
| jamielennox | gyee: sorry just saw your comment cause there was no +/- | 02:08 |
| jamielennox | gyee: why would someone want to develop a custom session? | 02:08 |
| jamielennox | gyee: i'm fine if they do but what for? | 02:08 |
| *** kgriffs is now known as kgriffs_afk | 02:08 | |
| ayoung | jamielennox, he wants a custome everything. It is the HP way | 02:08 |
| *** IanGovett has quit IRC | 02:09 | |
| *** ijw has joined #openstack-dev | 02:09 | |
| jamielennox | ayoung: i get the custom auth_plugin but that is well defined, the session is just a transport layer if there is something wrong with that then it is a bug i don't see why someone would want to reimplement it | 02:09 |
| bknudson | ERROR: openstackclient.shell Exception raised: (pbr 0.5.23.37.g4480343 (/opt/stack/pbr), Requirement.parse('pbr>=0.6,<1.0')) | 02:09 |
| bknudson | what do I do about that? | 02:10 |
| ayoung | bknudson, in your venv? | 02:10 |
| ayoung | activate it and pip upgrade pbr | 02:10 |
| bknudson | ayoung: no, not in this case... starting devstack | 02:10 |
| ayoung | bknudson, if its been a while since you ran devstack on this machine, maybe the pbr repo has moved ahead? Update the git repo> | 02:11 |
| bknudson | Can't uninstall 'pbr'. No files were found to uninstall. | 02:11 |
| jamielennox | ayoung: so anyway yes there is a v3 plugin and it handles all forms of v3 auth - password, token, oauth whatever | 02:11 |
| bknudson | -e git+https://github.com/openstack-dev/pbr.git@44803433a7da66b5e7404806290237469f07fd5d#egg=pbr-master | 02:12 |
| bknudson | that's the output of pip freeze | grep pbr | 02:12 |
| jamielennox | ayoung: i'd be willing to look at subclassing v3 into password/token/whatever - i'm not sure if thats a win or not | 02:12 |
| *** ijw has quit IRC | 02:12 | |
| ayoung | jamielennox, it would help if we got the code inmalingered. | 02:13 |
| ayoung | heh | 02:13 |
| ayoung | Its been malingering | 02:13 |
| ayoung | we're a little too detail oriented. Usually that is OK, but sometimes you need to ship. | 02:13 |
| *** emagana has quit IRC | 02:13 | |
| bknudson | ok, I do have an /opt/stack/pbr... so I git pull and setup.py install | 02:14 |
| *** yamahata has quit IRC | 02:14 | |
| bknudson | seems to work | 02:14 |
| ayoung | bknudson, is your +2 on https://review.openstack.org/#/c/61247/9 still good? Can I pull the trigger on it? | 02:15 |
| *** unicell has joined #openstack-dev | 02:15 | |
| *** jasondotstar has quit IRC | 02:15 | |
| bknudson | ayoung: I'll take a quick look and approve. | 02:16 |
| ayoung | bknudson, thanks. The important one is the one that follows it, and gyee and I have both +2ed it | 02:16 |
| dstanek | ayoung: thinking about notification of disabled entities - is this for internal use or are we advertising the capability outside of keystone? | 02:16 |
| ayoung | dstanek, I need for internal. External is a different conversation, and to be blunt, I don't care | 02:16 |
| dolphm | dstanek: they're emitted from keystone | 02:17 |
| ayoung | I can see arguments either way, but Federation is going to make it moot | 02:17 |
| ayoung | or Mute | 02:17 |
| dolphm | ayoung: you should use internal callbacks instead | 02:17 |
| *** gokrokve has joined #openstack-dev | 02:17 | |
| *** comay has quit IRC | 02:18 | |
| * ayoung goes and soaks his head | 02:18 | |
| dstanek | ayoung: i'm definitely not for it; i'd rather wait so that we can adjust as we learn more; i'm thinking of your question about a notification for changing passwords | 02:18 |
| ayoung | Change PW is different from user disable. | 02:18 |
| ayoung | But disable events...should go out to the world | 02:18 |
| ayoung | they are more important than update events | 02:19 |
| *** jasondotstar has joined #openstack-dev | 02:19 | |
| *** lbragstad has joined #openstack-dev | 02:19 | |
| *** sarob_ has quit IRC | 02:19 | |
| dolphm | updates events go out on disable already | 02:19 |
| *** sarob has joined #openstack-dev | 02:19 | |
| dolphm | i've been very hesitant on the redundant notifications | 02:20 |
| ayoung | dolphm, yes, but the end users don't have enough detail to distinguish between them | 02:20 |
| ayoung | I split them | 02:20 |
| dolphm | internal callsbacks seems like a sufficient baby step | 02:20 |
| ayoung | the notifications are either/or | 02:20 |
| dolphm | so you don't emit notifications on disable anymore? | 02:21 |
| dolphm | err.. | 02:21 |
| dolphm | you don't emit update notifications on disable anymore? | 02:21 |
| ayoung | dolphm, that is correct. | 02:21 |
| dolphm | so if anyone was subscribed to update notifications and only cared about disable events, they won't get them anymore? :-/ | 02:22 |
| ayoung | Yep | 02:22 |
| dolphm | what if disable is only part of the overall PATCH? i.e. name got updated as well | 02:22 |
| jamielennox | as a side append to what i was saying with notifications earlier the problem is they essentially are part of API now - people will start to rely on them | 02:22 |
| *** pradeep has joined #openstack-dev | 02:22 | |
| dstanek | can an update disable a user and change other stuff in the entity? | 02:22 |
| dstanek | dolphm: yes, exactly | 02:23 |
| ayoung | technically, yes, and only the disable would go through | 02:23 |
| jamielennox | can we just put a flag public=False on the notify callback which does the internal callback without the ampq? | 02:23 |
| ayoung | but...I still think that this is more correct | 02:23 |
| dolphm | jamielennox: ++ | 02:23 |
| jamielennox | we can start putting these things on the wire when we have a plan for this stuff | 02:23 |
| ayoung | disable is only an update due to implementation | 02:23 |
| ayoung | I think it is the disable evetns that the remote systems are most concerned about. If a user or project gets disabled, they want to disable resources | 02:24 |
| *** sarob has quit IRC | 02:24 | |
| dolphm | ayoung: you don't have an immediate use case for emitting disable events beyond keystone? so why not use internal callbacks | 02:24 |
| jamielennox | ayoung: user disable is a tough thing to commit to when we are moving away from having control over users | 02:24 |
| ayoung | Right now..I am guessing they have to do a callback to see status upon an update to see if it is disabled | 02:24 |
| jamielennox | (eg federation) | 02:25 |
| ayoung | hmmm...so roll back to the double notification approach, and tag the "disable" event as internal only? | 02:25 |
| jamielennox | ayoung: i think we tag as much as possible as internal only for the time being | 02:26 |
| ayoung | Do we have a mechanism for internal callbacks? Does notify do internal only? | 02:26 |
| dolphm | ayoung: ++ that seems like the least controversial solution for icehouse, without having to worry about backwards compatibility etc | 02:26 |
| dolphm | ayoung: yes and yes | 02:26 |
| *** markvoelker1 has joined #openstack-dev | 02:26 | |
| jamielennox | ayoung: it does both, but it would be a really easy thing to skip the ampq part if not public | 02:26 |
| dolphm | ayoung: https://blueprints.launchpad.net/keystone/+spec/internal-callbacks | 02:27 |
| *** yamahata has joined #openstack-dev | 02:27 | |
| jamielennox | dolphm: just that the implementation of that has squeezed in with the ampq stuff so at the moment you don't have control over just internal | 02:28 |
| *** sarob has joined #openstack-dev | 02:28 | |
| *** sarob has quit IRC | 02:28 | |
| *** primemin1sterp has joined #openstack-dev | 02:28 | |
| *** sarob has joined #openstack-dev | 02:28 | |
| jamielennox | anyway side question from something i had dropped earlier and we appear to have a quorum still awake | 02:29 |
| ayoung | Yeah...so how would I tag it as an internal only callback? | 02:29 |
| dolphm | jamielennox: ooooooh, i see what you mean | 02:29 |
| dolphm | jamielennox: yeah, we should add an internal_only kwarg :-/ | 02:29 |
| jamielennox | dolphm, ayoung, dstanek, bknudson: does having auth plugins like V3Password, V3Token etc make sense as opposed to a similar state as now where you have just V3Auth(username='', password='', token='') | 02:30 |
| dolphm | i remember talking to morgan about this now; didn't see this as a downside at the time | 02:30 |
| jamielennox | dolphm: public=False | 02:30 |
| dolphm | jamielennox: ++ | 02:30 |
| jamielennox | ayoung: it would be simple to add that param to the wrapper | 02:30 |
| *** achampio1 has joined #openstack-dev | 02:30 | |
| *** primeministerp has quit IRC | 02:31 | |
| *** erkules_ has joined #openstack-dev | 02:31 | |
| dstanek | jamielennox: i like the separate classes if there is different logic for each one that should be encapsulated | 02:31 |
| jamielennox | dolphm: was the ++ for the plugins or for the public=False? | 02:32 |
| jamielennox | dstanek: the main upside i see is adding new auth types to v3 and not extending that list | 02:32 |
| dolphm | jamielennox: oh, public | 02:32 |
| *** rkukura has joined #openstack-dev | 02:32 | |
| *** yamahata has quit IRC | 02:32 | |
| jamielennox | the main downside i see is when you have everything in a CONF or something where all you really want is a plugin that works | 02:32 |
| dolphm | jamielennox: and yes, i'd like them to be separate plugins | 02:32 |
| *** achampion has quit IRC | 02:33 | |
| bknudson | jamielennox: you instantiate one of these plugins and pass it in to the client? | 02:33 |
| jamielennox | ayoung: completely understand how you feel when you think you're almost done and things get changed | 02:33 |
| ayoung | does def _send_notification(operation, resource_type, resource_id, host=None) have to be internal? | 02:33 |
| jamielennox | bknudson: yes | 02:33 |
| jamielennox | ahh, | 02:33 |
| ayoung | jamielennox, heh | 02:33 |
| dolphm | jamielennox: i could see v3password extending v3token or something though, to manage refreshing? | 02:33 |
| jamielennox | bknudson: you pass it to the session | 02:33 |
| *** erkules has quit IRC | 02:33 | |
| jamielennox | dolphm: v3 token i was thinking auth_url/token | 02:33 |
| * ayoung is used to it by now. And we are better off catching these things before commit. | 02:34 | |
| jamielennox | they will all need to manage tokens | 02:34 |
| bknudson | ayoung: _send_notification doesn't have to be internal | 02:34 |
| bknudson | the wrappers are there are just because it was a convenient way to implement it... | 02:34 |
| bknudson | if it doesn't match how we want to use it now then there's no need to be stuck with just the wrappers. | 02:35 |
| jamielennox | ayoung: you're not making the function internal - just the emitted event | 02:35 |
| ayoung | jamielennox, so https://github.com/openstack/keystone/blob/master/keystone/notifications.py#L155 would be conditional on if (public): ? | 02:35 |
| jamielennox | bknudson: ++ i was playing with it recently and i think the wrapper is useful but not the only way | 02:35 |
| jamielennox | ayoung: right, indent that whole block with if public: | 02:36 |
| ayoung | jamielennox, yeah, but the decorators are kindof awkward for firing an event from the middle of a function. I'd just as soon call it explicitly | 02:36 |
| bknudson | jamielennox: so can I pass multiple auth plugins ... maybe pass both v3password and v3token? Not sure what it should do. | 02:36 |
| bknudson | maybe a v3password and a v2password and it figures out which one to use based on the endpoint? | 02:36 |
| jamielennox | ayoung: that's what i mean - we will need a way of doing a manual notify | 02:36 |
| jamielennox | bknudson: i have a scenario for dealing with multiple plugins in mind | 02:37 |
| jamielennox | we're not there yet | 02:37 |
| *** rtheis has quit IRC | 02:37 | |
| jamielennox | but it can/will work | 02:37 |
| bknudson | it would be sad to have the api version in the plugin. | 02:37 |
| jamielennox | bknudson: also endpoint are managed by the plugins (because endpoints come from the service catalog) | 02:37 |
| *** tsekiyama has joined #openstack-dev | 02:38 | |
| dolphm | bknudson: was just wondering if that was necessary as well... | 02:38 |
| *** achampion has joined #openstack-dev | 02:38 | |
| bknudson | jamielennox: wait, I thought with token auth you wouldn't have a catalog? | 02:38 |
| *** vkmc has quit IRC | 02:38 | |
| jamielennox | bknudson: there are two forms of token auth | 02:38 |
| jamielennox | endpoint/token is when you always use that endpoint/token | 02:39 |
| jamielennox | auth_url/token is when you rescope a token | 02:39 |
| *** amcrn has quit IRC | 02:39 | |
| jamielennox | or for whatever reason you get a token from a tokne | 02:39 |
| jamielennox | auth_url/token will give you a service catalog | 02:39 |
| jamielennox | endpoint/token won't - but it doesn't matter cause you are always using the same endpoint | 02:40 |
| *** achampio1 has quit IRC | 02:40 | |
| dolphm | +1 to all of the above | 02:40 |
| jamielennox | for future contemplation the way i was thinking of dealing with multiple auth plugins is to name them. So if you rescope a token you can have 'default' and 'project_scoped' for example | 02:41 |
| jamielennox | then when you create a client with that session you can say use auth='project_scoped' and it will tell the session which plugin it should use for requests | 02:42 |
| ayoung | jamielennox, I just +2ed Auth PLugins on top of bknudson approving the session patch. | 02:42 |
| jamielennox | i've still got to get even close to that though | 02:42 |
| *** tsekiyama has quit IRC | 02:42 | |
| ayoung | We're going to have a pretty cool client next release.... | 02:42 |
| jamielennox | ayoung: no - our client is ugly as sin, but 2.0 is shaping up nicely | 02:43 |
| ayoung | Heh | 02:43 |
| jamielennox | alright - guess now i need to rewrite my v2/v3 token plugins | 02:43 |
| *** ewindisch has quit IRC | 02:44 | |
| jamielennox | i'll try and run that one past dtroyer first | 02:44 |
| ayoung | jamielennox, so https://review.openstack.org/#/c/68007/4/keystoneclient/auth/identity/v3.py is your current approach, what are you going to do instead> | 02:44 |
| jamielennox | ayoung: 90% of that is common | 02:44 |
| jamielennox | but subclass the actual auth method | 02:44 |
| jamielennox | so a UsernamePassword(Auth) | 02:44 |
| ayoung | so a base v3 auth plugin ...I assume password will come out of the init param list? | 02:45 |
| jamielennox | and a UsernameToken(Auth) | 02:45 |
| ayoung | as will token | 02:45 |
| jamielennox | i'm not sure what becomes of user_id | 02:45 |
| *** markvoelker1 has quit IRC | 02:45 | |
| jamielennox | egh, user_domain_name etc will all have to go with username | 02:45 |
| jamielennox | this is why i think i opted to keep it how it was | 02:46 |
| *** nati_uen_ has quit IRC | 02:46 | |
| ayoung | jamielennox, yeah, I was having the same thoughts on the revocation. user: id, name, domain_id} is much cleaner than separate vars | 02:46 |
| jamielennox | so scoping and trust will stay as the base and the rest will go to subclasses | 02:46 |
| dolphm | jamielennox: do you think the auth plugins need to bother with exposing auth / scoping by ID's instead of just names? | 02:47 |
| ayoung | trust should somehow be under scope... | 02:47 |
| *** anniec has quit IRC | 02:47 | |
| jamielennox | dolphm: the only thing i can think of is that the CLI still allows that | 02:47 |
| jamielennox | and i think OSC as well | 02:47 |
| *** otherwiseguy has quit IRC | 02:48 | |
| jamielennox | dolphm, ayoung and bknudson: you've all been relatively quiet on the actual plugins - if you can see a way to clean up the parameters that we take for auth then now is the best time to speak up | 02:48 |
| jamielennox | dstanek: ^^ | 02:48 |
| bknudson | jamielennox: was looking at https://review.openstack.org/#/c/68007/4/keystoneclient/auth/identity/v3.py and it looks good to me. | 02:49 |
| jamielennox | bknudson: yea, i feel i'm biased at this point | 02:49 |
| jamielennox | i'm so used to parameters like user_domain_id that it makes sense | 02:49 |
| ayoung | jamielennox, how is the plugin going to be selected when run from the CLI? | 02:49 |
| *** markvoelker1 has joined #openstack-dev | 02:50 | |
| bknudson | seems like it would be better to complain about the scope conflict on construction rather than on get_auth_ref | 02:50 |
| *** markvoelker1 has quit IRC | 02:50 | |
| ayoung | I assume that is OS_PASSWORD is set, we get the password plugin until we have a token, and then we use the token plugin, right? | 02:50 |
| ayoung | and to do Kerberos or client auth, it will be an env var or CLI option? | 02:50 |
| jamielennox | ayoung: so auth discoverability is somewhat harder than client discoverabiility | 02:50 |
| bknudson | and I don't think the client should be raising AuthorizationFailure ... that should be coming from the server response only. | 02:51 |
| dstanek | jamielennox: looking at the review now | 02:51 |
| jamielennox | ayoung: this is why another reason i kept it close to the current layout | 02:51 |
| ayoung | jamielennox, simple things should be simple, and hard things should be possible | 02:51 |
| ayoung | what if it is just OS_AUTH_PLUGIN=<name> | 02:51 |
| jamielennox | it's really easy to just pass through all the current client params to just the v3.Auth method and nothing will change | 02:52 |
| ayoung | and we have a registry of names. If it isn't set it defaults to ...v2? | 02:52 |
| ayoung | v3? | 02:52 |
| dstanek | jamielennox: is that what you wanted to break up in to separate classes? | 02:52 |
| jamielennox | ayoung: heh, that was when someone was optimistic about a v3 client | 02:52 |
| jamielennox | auth version is ignored | 02:52 |
| dstanek | jamielennox: the Auth class from v3.py | 02:52 |
| jamielennox | silently i think | 02:53 |
| jamielennox | dstanek: yes, does it make sense to seperate a UsernamePasswordAuth from other types of auth like Oauth | 02:53 |
| ayoung | I'd prefer it if the param list for an Auth plugin specified what it requires. Parameters that are necessary for the plugin to work should be required params, and others optional, and only if they will be used | 02:53 |
| jamielennox | (OAuth is the one that's been bugging at me for a while now - it doesn't quite fit) | 02:53 |
| *** gordc has quit IRC | 02:54 | |
| jamielennox | ayoung: so the current approach has very much been give everything and then check the combinations | 02:54 |
| *** tsekiyama has joined #openstack-dev | 02:54 | |
| ayoung | We can do it using an array of parameter names, and have each method specify which list, and then have a common base list or something | 02:54 |
| ayoung | right. And I personally want passwords to die | 02:54 |
| *** anniec has joined #openstack-dev | 02:55 | |
| jamielennox | ayoung: i've been thinking about arrays of param names - if it works for the CLI to do that i'm happy to - but i think i would like to maintain some seperation between the plugins and the CLI | 02:55 |
| dstanek | jamielennox: i would definitely vote for separate classes that have smaller param lists | 02:55 |
| jamielennox | it's the CLIs job to figure out how to instantiate plugins | 02:55 |
| ayoung | ++ | 02:55 |
| jamielennox | dstanek: that's the way most of this has been going | 02:55 |
| *** smurugesan has joined #openstack-dev | 02:55 | |
| ayoung | plugin.required_params and .optional_params | 02:56 |
| jamielennox | as i said i'm so used to this stuff now i'm biased | 02:56 |
| dstanek | jamielennox: are all of the Auth param really optional? | 02:56 |
| jamielennox | dstanek: legacy of the old client | 02:56 |
| bknudson | will need some kind of plugin factory | 02:56 |
| jamielennox | because there are so many ways to mix and match options everything is optional and then you pick what you need and error if something is missing | 02:56 |
| jamielennox | bknudson: right, it was easy before - now we need some way of distinguishing plugins | 02:57 |
| jamielennox | and it would be good to have 3rd party plugins in that | 02:57 |
| bknudson | like kerberos or ssl client cert or something | 02:57 |
| jamielennox | but these are things i've felt that are not quite core to the actual plugin process | 02:57 |
| bknudson | or http basic auth | 02:57 |
| jamielennox | if we could define the interface to a plugin then we can go through that transformation later without problem | 02:58 |
| *** marcoemorais has quit IRC | 02:58 | |
| jamielennox | basic auth counts, kerberos and SSL only somewhat | 02:58 |
| jamielennox | auth plugins only deal with the HTTP side of authenticating | 02:59 |
| jamielennox | kerberos and SSL are transport parameters (even though they can/are used for auth) | 02:59 |
| ayoung | basic-auth! | 02:59 |
| jamielennox | hmm | 02:59 |
| ayoung | I have code for that somewhere | 02:59 |
| *** kgriffs_afk is now known as kgriffs | 02:59 | |
| jamielennox | crap - can i do ssl from a plugin? | 02:59 |
| jamielennox | uggh, it's possible i *might* be able to run ssl and kerberos from a plugin ... | 03:00 |
| jamielennox | that's difficult and may require a new plugin hook | 03:00 |
| jamielennox | but not impossible and probably not that ugly | 03:00 |
| jamielennox | with both kerberos and SSL plugins though you need to still actually have something to put in a token right? | 03:01 |
| bknudson | jamielennox: who knows how it's going to work with federation. | 03:02 |
| *** markwash has quit IRC | 03:03 | |
| bknudson | kerberos and SSL don't have all the info you need for a regular auth request -- the scope. | 03:03 |
| jamielennox | bknudson: i'm just wondering if you set either kerberos or SSL certs via an auth plugin what do you expect then? are these used for every request or just auth? | 03:03 |
| bknudson | we should have the scope in the url... /v3/auth/tokens/project/<id> | 03:04 |
| *** evilstephen has joined #openstack-dev | 03:04 | |
| jamielennox | bknudson: right in the tests i've done with kerberos/ssl it was alway in addition to regular authentication | 03:04 |
| bknudson | jamielennox: good question... once you have a token you don't need to keep presenting the client cert everywhere. | 03:04 |
| *** angdraug has quit IRC | 03:05 | |
| jamielennox | bknudson: but then do i have to put the option of SSL certs and kerberos into the base of every auth plugin? | 03:05 |
| bknudson | that's where the question of multiple plugins comes in. | 03:05 |
| jamielennox | bknudson: so i don't think we should have multiple cooperative plugins | 03:06 |
| *** gokrokve has quit IRC | 03:06 | |
| *** dkranz has joined #openstack-dev | 03:06 | |
| jamielennox | everything i've done has been on the impression that 1 plugin is 1 token | 03:06 |
| jamielennox | you can then ask things of it that are relevant to that token | 03:06 |
| bknudson | jamielennox: you can do re-auth with the username/password plugin, though? | 03:07 |
| *** fandi has joined #openstack-dev | 03:07 | |
| jamielennox | bknudson: yes, it was decided that that should be handled internally to the plugin | 03:07 |
| bknudson | hopefully you can tell when your token is close to expiring, or it did expire. | 03:07 |
| jamielennox | so when i ask for a token the plugin has the option to refresh it | 03:07 |
| bknudson | jamielennox: the plugin stores the token? is that in BaseIdentityPlugin | 03:09 |
| jamielennox | yes | 03:09 |
| *** kgriffs is now known as kgriffs_afk | 03:09 | |
| *** paragan has joined #openstack-dev | 03:10 | |
| *** markmcclain has joined #openstack-dev | 03:10 | |
| *** CaptTofu has quit IRC | 03:11 | |
| *** relaxdiego has joined #openstack-dev | 03:14 | |
| *** anniec has quit IRC | 03:14 | |
| jamielennox | bknudson: re https://review.openstack.org/#/c/70902/2 | 03:18 |
| jamielennox | i thought we used 300 for / always if there is only one option i should return 200? | 03:18 |
| *** primemin1sterp has quit IRC | 03:20 | |
| bknudson | jamielennox: it doesn't make sense to me to return 300 multiple choice when there's only one choice | 03:22 |
| bknudson | a regular redirect makes more sense to me. | 03:23 |
| *** caleb_ has joined #openstack-dev | 03:23 | |
| jamielennox | bknudson: a redirect for / -> v1 ? that seems dangerous for later | 03:23 |
| bknudson | jamielennox: I assume it's going to be documented in the api reference. | 03:24 |
| *** alexpilotti has joined #openstack-dev | 03:24 | |
| jamielennox | right but you consult / for the available versions and then you get the link to your api versoin | 03:24 |
| bknudson | jamielennox: I guess 200 makes sense then. | 03:25 |
| bknudson | I thought that this was implementing some spec... what's the spec say? | 03:26 |
| *** kgriffs_afk is now known as kgriffs | 03:26 | |
| *** yamahata has joined #openstack-dev | 03:26 | |
| bknudson | jamielennox: this should all be defined in https://github.com/openstack/identity-api/blob/master/openstack-identity-api/v3/src/markdown/identity-api-v3-os-kds-ext.md | 03:27 |
| *** alop has joined #openstack-dev | 03:28 | |
| *** arnaud__ has quit IRC | 03:29 | |
| *** arnaud___ has quit IRC | 03:29 | |
| jamielennox | bknudson: similar to i guess v3 i only define the actual api interaction for v1 | 03:29 |
| jamielennox | also regarding discovery http://paste.openstack.org/show/62947/ | 03:30 |
| jamielennox | thats a sample of some of what's used | 03:30 |
| *** caleb_ has quit IRC | 03:30 | |
| bknudson | what a mess | 03:30 |
| bknudson | what is kds going to use? | 03:30 |
| *** yamahata has quit IRC | 03:31 | |
| jamielennox | bknudson: this is what i used to generate them: http://paste.openstack.org/show/62949/ | 03:31 |
| *** mlavalle has quit IRC | 03:31 | |
| jamielennox | i've been meaning to run it on a full devstack to get a wider sample but i seem to not have access to our OS instance | 03:31 |
| bknudson | jamielennox: do they all return 300? | 03:31 |
| jamielennox | bknudson: didn't print that | 03:31 |
| *** tchaypo is now known as jamezpolley | 03:32 | |
| bknudson | v2.0 is beta? | 03:32 |
| *** jamezpolley is now known as tchaypo | 03:32 | |
| bknudson | I think this is all the more reason to document what you plan to do and get agreement on that | 03:33 |
| jamielennox | i have a feeling it's an old instance | 03:33 |
| *** RajeshMohan has quit IRC | 03:34 | |
| ayoung | jamie, if there is no scope, and you do REMOTE_USER auth of any sort, the token is an unscoped token. They can always trade that for a scoped token | 03:35 |
| *** RajeshMohan has joined #openstack-dev | 03:35 | |
| jamielennox | ayoung: sure | 03:36 |
| *** kgriffs is now known as kgriffs_afk | 03:36 | |
| ayoung | I'd probably make the scoping parameters on the CLI/env vars, and make a scope object that the auth plugins can accept that gets apssed through verbatim to the token request | 03:36 |
| *** doug_shelley66 has quit IRC | 03:38 | |
| jamielennox | ayoung: the auth plugins are going to need to deal with both, how the user of the plugin gets that information isn't up to the plugin right? | 03:38 |
| ayoung | correct. something builds the scope and then the auth plugin just accepts a scope object | 03:38 |
| *** gokrokve has joined #openstack-dev | 03:38 | |
| ayoung | heh, its a pipeline, just the mirror of what we want to build on the token side in the server | 03:38 |
| *** carl_baldwin has joined #openstack-dev | 03:39 | |
| *** anniec has joined #openstack-dev | 03:39 | |
| *** gokrokve_ has joined #openstack-dev | 03:40 | |
| *** sarob has quit IRC | 03:40 | |
| *** sarob has joined #openstack-dev | 03:41 | |
| *** baoli has quit IRC | 03:42 | |
| *** gokrokve has quit IRC | 03:43 | |
| *** carl_baldwin has quit IRC | 03:43 | |
| ayoung | jamielennox, is anything holding up https://review.openstack.org/#/c/70664/ as I think that is the key one for RDO | 03:43 |
| *** evilstephen has quit IRC | 03:43 | |
| *** gokrokve_ has quit IRC | 03:44 | |
| *** bswartz has joined #openstack-dev | 03:45 | |
| jamielennox | ayoung: there are 3 patches in front of it | 03:45 |
| jamielennox | ayoung: damn didn't see that they need rebasing | 03:45 |
| jamielennox | ayoung: start: https://review.openstack.org/#/c/70661/2 | 03:46 |
| *** sarob has quit IRC | 03:46 | |
| ayoung | re-approved | 03:46 |
| jamielennox | hmm lbragstad pushed an updated one - i missed that | 03:46 |
| jamielennox | grr | 03:46 |
| jamielennox | :) | 03:47 |
| lbragstad | jamielennox: oh the doc change? | 03:47 |
| *** sarob has joined #openstack-dev | 03:47 | |
| lbragstad | I had dependent on yours? | 03:47 |
| jamielennox | lbragstad: yea, it recommited the first patch so the others lost there dep | 03:47 |
| *** buzztroll has quit IRC | 03:47 | |
| lbragstad | ahhh | 03:47 |
| ayoung | jamielennox, rebase https://review.openstack.org/#/c/71044/ as well then | 03:48 |
| ayoung | PITA | 03:48 |
| *** sarob has quit IRC | 03:49 | |
| jamielennox | ayoung: yep did the first one, because that one belongs to jay i have to do it manually | 03:49 |
| ayoung | Ok my approval spree is over. I'll look back in come the morning. | 03:49 |
| *** sarob has joined #openstack-dev | 03:49 | |
| jamielennox | ayoung: so the reason the last one is failing tests is cause oslo.messaging defaults to using the rabbit driver | 03:49 |
| jamielennox | which defaults to a dependency on kombu | 03:50 |
| jamielennox | does that mean that keystone needs a dependency on kombu - i can't see any other choice but it seems unnecessary | 03:50 |
| ayoung | AHHHHHHHH | 03:51 |
| *** yaguang has joined #openstack-dev | 03:52 | |
| *** harlowja is now known as harlowja_away | 03:53 | |
| *** carl_baldwin has joined #openstack-dev | 03:54 | |
| *** sarob has quit IRC | 03:54 | |
| *** aditirav has joined #openstack-dev | 03:54 | |
| jamielennox | rage quit | 03:55 |
| jamielennox | bknudson: you'll like this - a larger sample: http://paste.openstack.org/show/62955/ | 03:56 |
| jamielennox | so there is precendent with heat for using a 300 with only one version | 03:56 |
| bknudson | it's just totally random... | 03:57 |
| bknudson | should just return rand() | 03:57 |
| jamielennox | and cinder has 2 versions and uses 200 | 03:57 |
| *** markmcclain1 has joined #openstack-dev | 03:57 | |
| *** markmcclain has quit IRC | 03:57 | |
| *** alex_klimov has joined #openstack-dev | 03:58 | |
| ayoung | what is that utility that gets you command history in aplications that don';t have it...I used it with oracles command line years ago... | 03:59 |
| StevenK | ayoung: readline? | 03:59 |
| jamielennox | bknudson: so i can use whatever i like? | 03:59 |
| ayoung | StevenK, nah, it is something that wraps another command prompt program. I though it was a two letter command | 03:59 |
| ayoung | fc I thin | 04:00 |
| *** edmund has quit IRC | 04:00 | |
| jamielennox | bknudson: most services don't have an updated_at - i think that ones kind of useless | 04:00 |
| bknudson | jamielennox: you can use whatever's documented | 04:00 |
| ayoung | nope | 04:00 |
| StevenK | ayoung: rlwrap | 04:00 |
| jamielennox | i thought other servies had media-types as well but it appears its just us | 04:00 |
| ayoung | StevenK, that sounds promising | 04:01 |
| *** morazi has quit IRC | 04:01 | |
| *** tiamar has joined #openstack-dev | 04:03 | |
| ayoung | StevenK, 31 packages of Perl now installing | 04:04 |
| StevenK | Haha | 04:04 |
| *** alex_klimov has quit IRC | 04:04 | |
| *** aveiga has quit IRC | 04:05 | |
| ayoung | StevenK, thanks...that was it | 04:06 |
| StevenK | ayoung: You're welcome | 04:06 |
| *** ytwu has joined #openstack-dev | 04:06 | |
| *** alexpilotti has quit IRC | 04:07 | |
| *** radix has left #openstack-dev | 04:08 | |
| *** marcoemorais has joined #openstack-dev | 04:08 | |
| *** marcoemorais has quit IRC | 04:08 | |
| *** ayoung is now known as ayoung-ZZZZZ | 04:09 | |
| *** radix has joined #openstack-dev | 04:09 | |
| *** xarses has joined #openstack-dev | 04:10 | |
| *** changbl has joined #openstack-dev | 04:13 | |
| *** yamahata has joined #openstack-dev | 04:17 | |
| *** dstanek has quit IRC | 04:19 | |
| *** dstanek has joined #openstack-dev | 04:19 | |
| *** evilstephen has joined #openstack-dev | 04:21 | |
| *** sarob has joined #openstack-dev | 04:21 | |
| *** carl_baldwin has quit IRC | 04:21 | |
| *** ayoung-ZZZZZ has quit IRC | 04:22 | |
| *** anniec has quit IRC | 04:22 | |
| *** evilstephen has quit IRC | 04:25 | |
| *** kgriffs_afk is now known as kgriffs | 04:27 | |
| *** mikeoutland has joined #openstack-dev | 04:27 | |
| *** bhuvan has joined #openstack-dev | 04:28 | |
| *** anniec has joined #openstack-dev | 04:30 | |
| *** bhuvan has quit IRC | 04:33 | |
| *** sarob has quit IRC | 04:34 | |
| *** sarob has joined #openstack-dev | 04:35 | |
| *** kgriffs is now known as kgriffs_afk | 04:36 | |
| *** halfie has joined #openstack-dev | 04:36 | |
| *** pablosan has quit IRC | 04:37 | |
| *** buzztroll has joined #openstack-dev | 04:38 | |
| *** evilstephen has joined #openstack-dev | 04:38 | |
| *** sarob has quit IRC | 04:39 | |
| *** sarob has joined #openstack-dev | 04:39 | |
| *** pablosan has joined #openstack-dev | 04:39 | |
| *** gokrokve has joined #openstack-dev | 04:39 | |
| *** Tross has joined #openstack-dev | 04:40 | |
| *** gokrokve_ has joined #openstack-dev | 04:41 | |
| *** Tross has left #openstack-dev | 04:42 | |
| *** buzztroll has quit IRC | 04:42 | |
| *** Ryan_Lane has quit IRC | 04:44 | |
| *** gokrokve has quit IRC | 04:44 | |
| *** thuc has joined #openstack-dev | 04:44 | |
| *** terrylhowe has quit IRC | 04:45 | |
| *** sweston_ has quit IRC | 04:46 | |
| *** thuc has quit IRC | 04:47 | |
| *** thuc has joined #openstack-dev | 04:48 | |
| *** tsekiyama has quit IRC | 04:48 | |
| *** pabelanger has left #openstack-dev | 04:49 | |
| *** markmcclain1 has quit IRC | 04:50 | |
| *** markmcclain has joined #openstack-dev | 04:50 | |
| markmcclain | sdague or mtreinish: around ? | 04:51 |
| *** thuc has quit IRC | 04:52 | |
| *** 23LAA4WAK has joined #openstack-dev | 04:53 | |
| *** 23LAA4WAK has quit IRC | 04:53 | |
| *** markmcclain1 has joined #openstack-dev | 04:54 | |
| *** markmcclain2 has joined #openstack-dev | 04:54 | |
| *** pcm_ has quit IRC | 04:55 | |
| *** armax has quit IRC | 04:55 | |
| *** markmcclain has quit IRC | 04:56 | |
| *** gcha has quit IRC | 04:58 | |
| *** markmcclain1 has quit IRC | 04:59 | |
| *** comay has joined #openstack-dev | 04:59 | |
| *** clayb|2 has quit IRC | 05:03 | |
| *** radsy has quit IRC | 05:03 | |
| *** amcrn has joined #openstack-dev | 05:04 | |
| *** killer_prince has joined #openstack-dev | 05:06 | |
| *** Ryan_Lane has joined #openstack-dev | 05:07 | |
| *** relaxdiego has quit IRC | 05:07 | |
| *** saju_m has joined #openstack-dev | 05:08 | |
| *** amcrn_ has joined #openstack-dev | 05:08 | |
| *** markmcclain2 has quit IRC | 05:10 | |
| *** amcrn has quit IRC | 05:11 | |
| *** gokrokve_ has quit IRC | 05:11 | |
| *** gokrokve has joined #openstack-dev | 05:11 | |
| *** CaptTofu has joined #openstack-dev | 05:12 | |
| *** gokrokve has quit IRC | 05:12 | |
| *** gokrokve has joined #openstack-dev | 05:12 | |
| *** gokrokve has quit IRC | 05:12 | |
| *** gokrokve has joined #openstack-dev | 05:13 | |
| *** gokrokve has quit IRC | 05:13 | |
| *** sarob has quit IRC | 05:14 | |
| *** sarob has joined #openstack-dev | 05:14 | |
| *** jcooley_ has quit IRC | 05:14 | |
| *** jcooley_ has joined #openstack-dev | 05:15 | |
| *** CaptTofu has quit IRC | 05:16 | |
| *** gokrokve has joined #openstack-dev | 05:18 | |
| *** gokrokve has quit IRC | 05:18 | |
| *** sarob has quit IRC | 05:19 | |
| *** ytwu has quit IRC | 05:21 | |
| *** alex_xu has joined #openstack-dev | 05:21 | |
| *** harlowja_at_home has joined #openstack-dev | 05:22 | |
| *** kushal has joined #openstack-dev | 05:22 | |
| *** nshaikh has joined #openstack-dev | 05:22 | |
| *** mikeoutland has quit IRC | 05:22 | |
| *** aditirav_ has joined #openstack-dev | 05:23 | |
| *** aditirav has quit IRC | 05:26 | |
| *** aditirav_ is now known as aditirav | 05:26 | |
| *** kgriffs_afk is now known as kgriffs | 05:27 | |
| *** aditirav has quit IRC | 05:28 | |
| *** aditirav has joined #openstack-dev | 05:28 | |
| *** kushal has quit IRC | 05:31 | |
| *** buzztroll has joined #openstack-dev | 05:32 | |
| *** rohitk has joined #openstack-dev | 05:34 | |
| *** stevemar has joined #openstack-dev | 05:35 | |
| *** gyee has quit IRC | 05:36 | |
| *** unicell has quit IRC | 05:37 | |
| *** kgriffs is now known as kgriffs_afk | 05:37 | |
| *** kushal has joined #openstack-dev | 05:39 | |
| *** kushal has joined #openstack-dev | 05:39 | |
| *** harlowja_at_home has quit IRC | 05:41 | |
| *** mriedem has quit IRC | 05:41 | |
| *** rohitk has quit IRC | 05:44 | |
| *** sarob has joined #openstack-dev | 05:45 | |
| *** rdas has joined #openstack-dev | 05:47 | |
| *** achampion has quit IRC | 05:47 | |
| *** pradeep1 has joined #openstack-dev | 05:50 | |
| *** pradeep has quit IRC | 05:52 | |
| *** gokrokve has joined #openstack-dev | 05:53 | |
| *** gokrokve has quit IRC | 05:58 | |
| *** mikeoutland has joined #openstack-dev | 05:58 | |
| *** rohitk has joined #openstack-dev | 06:00 | |
| *** sweston has joined #openstack-dev | 06:01 | |
| *** doug_shelley66 has joined #openstack-dev | 06:02 | |
| *** sarob has quit IRC | 06:07 | |
| *** neeti has joined #openstack-dev | 06:08 | |
| *** mikeoutland has quit IRC | 06:09 | |
| *** paragan has quit IRC | 06:13 | |
| *** achampion has joined #openstack-dev | 06:14 | |
| *** sarob has joined #openstack-dev | 06:14 | |
| *** paragan has joined #openstack-dev | 06:16 | |
| *** mrda is now known as mrda_away | 06:17 | |
| *** comay has quit IRC | 06:17 | |
| *** saju_m has quit IRC | 06:18 | |
| *** achampion has quit IRC | 06:18 | |
| *** tqtran has joined #openstack-dev | 06:20 | |
| *** sarob has quit IRC | 06:25 | |
| *** denis_makogon has joined #openstack-dev | 06:25 | |
| *** sarob has joined #openstack-dev | 06:25 | |
| *** DinaBelova_ is now known as DinaBelova | 06:26 | |
| *** kgriffs_afk is now known as kgriffs | 06:28 | |
| *** neeti has quit IRC | 06:30 | |
| *** sarob has quit IRC | 06:30 | |
| *** qs201 has joined #openstack-dev | 06:32 | |
| *** qs201 has quit IRC | 06:34 | |
| *** cfriesen has quit IRC | 06:35 | |
| *** cadenzajon has joined #openstack-dev | 06:37 | |
| *** jamespage_ has joined #openstack-dev | 06:37 | |
| *** jasondotstar has quit IRC | 06:38 | |
| *** kgriffs is now known as kgriffs_afk | 06:38 | |
| *** alex_klimov has joined #openstack-dev | 06:39 | |
| *** gokrokve has joined #openstack-dev | 06:39 | |
| *** amotoki has joined #openstack-dev | 06:40 | |
| *** smurugesan has quit IRC | 06:41 | |
| *** pradeep1 has quit IRC | 06:41 | |
| *** cadenzajon has quit IRC | 06:41 | |
| *** gokrokve has quit IRC | 06:44 | |
| *** pschaef has joined #openstack-dev | 06:46 | |
| *** anniec has quit IRC | 06:48 | |
| *** mikeoutland has joined #openstack-dev | 06:49 | |
| *** jprovazn has joined #openstack-dev | 06:50 | |
| *** tqtran has quit IRC | 06:51 | |
| *** bhuvan has joined #openstack-dev | 06:52 | |
| *** erkules_ is now known as erkules | 06:53 | |
| *** pcm_ has joined #openstack-dev | 06:56 | |
| *** stevemar has quit IRC | 06:56 | |
| *** neeti has joined #openstack-dev | 06:59 | |
| *** sarob has joined #openstack-dev | 07:00 | |
| *** pcm_ has quit IRC | 07:00 | |
| *** sarob has quit IRC | 07:01 | |
| *** pradeep has joined #openstack-dev | 07:01 | |
| *** sarob has joined #openstack-dev | 07:01 | |
| *** alex_klimov has quit IRC | 07:01 | |
| *** jcooley_ has quit IRC | 07:05 | |
| *** matsuhashi has quit IRC | 07:05 | |
| *** sarob has quit IRC | 07:06 | |
| *** jcooley_ has joined #openstack-dev | 07:06 | |
| *** MaxV has joined #openstack-dev | 07:06 | |
| *** jcooley_ has quit IRC | 07:08 | |
| *** vartom1111111117 has joined #openstack-dev | 07:09 | |
| *** CaptTofu has joined #openstack-dev | 07:13 | |
| *** NikitaKonovalov_ is now known as NikitaKonovalov | 07:13 | |
| *** lcheng_ has joined #openstack-dev | 07:14 | |
| *** sushils has quit IRC | 07:17 | |
| *** CaptTofu has quit IRC | 07:18 | |
| *** dkranz has quit IRC | 07:18 | |
| *** dkranz has joined #openstack-dev | 07:18 | |
| *** markwash has joined #openstack-dev | 07:20 | |
| *** NikitaKonovalov is now known as NikitaKonovalov_ | 07:20 | |
| *** matsuhashi has joined #openstack-dev | 07:21 | |
| *** rgerganov has joined #openstack-dev | 07:21 | |
| *** dstanek has quit IRC | 07:23 | |
| *** buzztroll has quit IRC | 07:25 | |
| *** buzztroll has joined #openstack-dev | 07:25 | |
| *** buzztroll has quit IRC | 07:25 | |
| *** Drankis has joined #openstack-dev | 07:26 | |
| *** buzztroll has joined #openstack-dev | 07:26 | |
| *** kgriffs_afk is now known as kgriffs | 07:29 | |
| *** dstufft_ has joined #openstack-dev | 07:29 | |
| *** amcrn_ is now known as amcrn | 07:30 | |
| *** jamieh has joined #openstack-dev | 07:31 | |
| *** ytwu has joined #openstack-dev | 07:31 | |
| *** dstufft has quit IRC | 07:32 | |
| *** saju_m has joined #openstack-dev | 07:32 | |
| *** jcooley_ has joined #openstack-dev | 07:34 | |
| *** yolanda has joined #openstack-dev | 07:34 | |
| *** rohitk has quit IRC | 07:34 | |
| *** Drankis has quit IRC | 07:37 | |
| *** comay has joined #openstack-dev | 07:38 | |
| *** kgriffs is now known as kgriffs_afk | 07:39 | |
| *** gokrokve has joined #openstack-dev | 07:39 | |
| *** jhesketh__ has quit IRC | 07:40 | |
| *** markwash has quit IRC | 07:40 | |
| *** AlexF has joined #openstack-dev | 07:41 | |
| *** bvandenh has joined #openstack-dev | 07:43 | |
| *** athomas has quit IRC | 07:43 | |
| *** gokrokve has quit IRC | 07:44 | |
| *** DinaBelova is now known as DinaBelova_ | 07:44 | |
| *** dkuffner has joined #openstack-dev | 07:45 | |
| *** MaxV has quit IRC | 07:45 | |
| *** markwash has joined #openstack-dev | 07:47 | |
| *** taps has quit IRC | 07:47 | |
| *** rohitk has joined #openstack-dev | 07:47 | |
| *** bauzas has joined #openstack-dev | 07:48 | |
| *** dstufft_ is now known as dstufft | 07:50 | |
| *** nosnos_ has joined #openstack-dev | 07:51 | |
| *** paragan has quit IRC | 07:52 | |
| *** NikitaKonovalov_ is now known as NikitaKonovalov | 07:52 | |
| *** kushal has quit IRC | 07:53 | |
| *** nosnos has quit IRC | 07:54 | |
| *** afazekas_ has joined #openstack-dev | 07:56 | |
| *** johnthetubaguy has joined #openstack-dev | 08:01 | |
| *** achampion has joined #openstack-dev | 08:03 | |
| *** johnthetubaguy has quit IRC | 08:03 | |
| *** johnthetubaguy has joined #openstack-dev | 08:03 | |
| *** denis_makogon has quit IRC | 08:04 | |
| *** smurugesan has joined #openstack-dev | 08:06 | |
| *** kushal has joined #openstack-dev | 08:07 | |
| *** gcha has joined #openstack-dev | 08:08 | |
| *** smurugesan has quit IRC | 08:10 | |
| *** matsuhashi has quit IRC | 08:10 | |
| *** sarob has joined #openstack-dev | 08:12 | |
| *** lari_ has quit IRC | 08:12 | |
| *** johnthetubaguy has quit IRC | 08:12 | |
| *** oro has joined #openstack-dev | 08:13 | |
| *** afazekas_ has quit IRC | 08:13 | |
| *** paragan has joined #openstack-dev | 08:13 | |
| *** Ryan_Lane has quit IRC | 08:14 | |
| *** lari_ has joined #openstack-dev | 08:14 | |
| *** jamespage_ has joined #openstack-dev | 08:15 | |
| *** sarob has quit IRC | 08:17 | |
| *** reed has joined #openstack-dev | 08:20 | |
| *** posito has joined #openstack-dev | 08:22 | |
| *** lcheng_ has quit IRC | 08:23 | |
| posito | Hello, can i apply manually a patch from review.openstack.org (devstack-gate is a bit overkill for my need). | 08:24 |
| *** jamespage_ has quit IRC | 08:24 | |
| *** xga has joined #openstack-dev | 08:25 | |
| *** xga_ has joined #openstack-dev | 08:25 | |
| rushiagr | posito: for submissions upstream, no, you cannot skip the process | 08:25 |
| *** martyntaylor has joined #openstack-dev | 08:25 | |
| posito | rushiagr: I am not trying to vote just understand when I have to do the cherry-pick in my flow | 08:25 |
| rushiagr | posito: sorry, I didn't get you | 08:26 |
| posito | rushiagr: regarding my message it's not clear. Let me try it again. I want to test (https://review.openstack.org/#/c/70835/) How should I apply the patch. | 08:28 |
| *** MaxV has joined #openstack-dev | 08:28 | |
| *** flaper87|afk is now known as flaper87 | 08:29 | |
| rushiagr | posito: oh, okay. Sorry, I misunderstood you | 08:29 |
| rushiagr | you can see a line starting with 'git fetch https://....' on the review page | 08:30 |
| *** kgriffs_afk is now known as kgriffs | 08:30 | |
| rushiagr | and a copy icon next to it | 08:30 |
| rushiagr | posito: so just copy it and paste it into your local repository if you want to apply that patch | 08:30 |
| posito | rushiagr: Yes... but when I will do a stack.sh those change will discarded unless i go in offline mode | 08:31 |
| rushiagr | posito: you are right | 08:31 |
| *** bauzas has quit IRC | 08:31 | |
| rushiagr | posito: for that, what you can do is: 1. set up devstack 2. apply this patch 3. restart the affected screens | 08:31 |
| rushiagr | posito: I'm not a compute expert so can't say what all services you need to restart inside the screen sessions | 08:32 |
| rushiagr | posito: i'll just restart all screens starting with 'n-' :) | 08:32 |
| rushiagr | posito: ask me if you need help regarding screens | 08:33 |
| *** mmagr has joined #openstack-dev | 08:35 | |
| posito | rushiagr:actually i am interested in neutron. So i should do is 1) run my stack normally. 2)go the location where stack is installed. do a git fetch 3) restart neutron | 08:35 |
| *** nacim has joined #openstack-dev | 08:36 | |
| rushiagr | posito: I am not completely sure, but the code affected by the change is only in nova, so you need to restart nova services only | 08:36 |
| posito | rushiagr: the review that I showed you earlier was just a sample. But in general is there any patch that will change more than one service ? | 08:38 |
| rushiagr | posito: I don't think so | 08:38 |
| *** sahid has joined #openstack-dev | 08:38 | |
| posito | rushiagr: also when I restart the service it will drop and recreate the mandatory database ? | 08:39 |
| rushiagr | posito: no | 08:39 |
| rushiagr | it won't as far as I know | 08:39 |
| *** gokrokve has joined #openstack-dev | 08:39 | |
| *** florentflament has joined #openstack-dev | 08:40 | |
| *** kgriffs is now known as kgriffs_afk | 08:40 | |
| posito | rushiagr: hum... will it create them or nothing at all | 08:40 |
| *** ndipanov_gone is now known as ndipanov | 08:41 | |
| rushiagr | posito: it will just restart the service, and won't recreate databases. However, restarting screens might affect database entries, possibly | 08:41 |
| posito | rushiagr: so maybe a better way would be to restart the screens | 08:41 |
| rushiagr | posito: right | 08:42 |
| *** gokrokve has quit IRC | 08:44 | |
| *** iartarisi has joined #openstack-dev | 08:46 | |
| posito | rushiagr: damnint i thought i could use rejoin-stack.sh | 08:47 |
| *** mrunge has joined #openstack-dev | 08:48 | |
| posito | rushiagr: the description is strange though (Restart openstack services after running stack.sh) | 08:48 |
| rushiagr | posito: yes, it restarts all services, if all services are not running. But I'm not sure if it works with some of the services killed and the others running | 08:49 |
| *** haomai___ has quit IRC | 08:49 | |
| posito | rushiagr: I could do a kill -9 on the neutron PID and then run this | 08:50 |
| *** haomaiwang has joined #openstack-dev | 08:50 | |
| *** sushils has joined #openstack-dev | 08:51 | |
| rushiagr | posito: is it throwing some error? Generally i've used rejoin stack only if all the devstack processes are not running | 08:51 |
| posito | rushiagr: i didn't try it it . | 08:53 |
| *** lcheng_ has joined #openstack-dev | 08:53 | |
| posito | rushiagr: i didn't try it yet. - I assume that it is is down, it will try to re-start the service | 08:53 |
| rushiagr | posito: or you can just go to that screen, kill the process with ctrl+C, press up arrow to see the last statement executed, and execute it | 08:54 |
| posito | rushiagr: I would like to try to do this with a script | 08:55 |
| *** ygbo has joined #openstack-dev | 08:55 | |
| rushiagr | posito: oh | 08:55 |
| posito | rushiagr: the devstack-gate, seems to be really great but a bit overkill with all of the customization that i need to do before | 08:56 |
| *** DinaBelova_ is now known as DinaBelova | 08:57 | |
| *** yassine has joined #openstack-dev | 08:58 | |
| *** cnesa has joined #openstack-dev | 08:58 | |
| *** nkinder has joined #openstack-dev | 08:59 | |
| *** jpich has joined #openstack-dev | 09:00 | |
| *** zzelle has joined #openstack-dev | 09:01 | |
| *** AlexF has quit IRC | 09:02 | |
| *** markmc has joined #openstack-dev | 09:04 | |
| *** xqueralt has joined #openstack-dev | 09:05 | |
| *** julienvey has joined #openstack-dev | 09:06 | |
| *** sarob has joined #openstack-dev | 09:07 | |
| *** bauzas has joined #openstack-dev | 09:07 | |
| *** Ryan_Lane has joined #openstack-dev | 09:07 | |
| *** amerine_ has joined #openstack-dev | 09:09 | |
| *** DinaBelova is now known as DinaBelova_ | 09:09 | |
| *** skudriashev has joined #openstack-dev | 09:10 | |
| *** amerine has quit IRC | 09:11 | |
| *** posito has quit IRC | 09:11 | |
| *** sarob has quit IRC | 09:11 | |
| *** CaptTofu has joined #openstack-dev | 09:14 | |
| *** stannie has joined #openstack-dev | 09:14 | |
| *** marun has quit IRC | 09:16 | |
| *** mflobo has quit IRC | 09:17 | |
| *** safchain has joined #openstack-dev | 09:17 | |
| *** Ryan_Lane has quit IRC | 09:17 | |
| *** CaptTofu has quit IRC | 09:18 | |
| *** eglynn has joined #openstack-dev | 09:19 | |
| *** rdas has quit IRC | 09:22 | |
| *** derekh has joined #openstack-dev | 09:22 | |
| *** bhuvan has quit IRC | 09:22 | |
| *** marekd|away is now known as marekd | 09:22 | |
| *** yolanda has quit IRC | 09:26 | |
| *** yolanda has joined #openstack-dev | 09:26 | |
| *** sweston has quit IRC | 09:27 | |
| *** DinaBelova_ is now known as DinaBelova | 09:27 | |
| *** lucasagomes has joined #openstack-dev | 09:30 | |
| *** kgriffs_afk is now known as kgriffs | 09:31 | |
| *** lcheng_ has quit IRC | 09:32 | |
| *** giulivo has joined #openstack-dev | 09:35 | |
| *** gilliard has joined #openstack-dev | 09:36 | |
| *** Oneiroi has joined #openstack-dev | 09:39 | |
| *** kgriffs is now known as kgriffs_afk | 09:40 | |
| *** nkinder has quit IRC | 09:41 | |
| *** mkollaro has joined #openstack-dev | 09:41 | |
| *** lari_ has quit IRC | 09:41 | |
| *** nosnos_ has quit IRC | 09:43 | |
| *** lari_ has joined #openstack-dev | 09:45 | |
| *** max_lobur_afk is now known as max_lobur | 09:46 | |
| *** sushils has quit IRC | 09:49 | |
| *** YorikSar has joined #openstack-dev | 09:51 | |
| *** bhuvan has joined #openstack-dev | 09:51 | |
| *** Ryan_Lane has joined #openstack-dev | 09:54 | |
| *** mkollaro1 has joined #openstack-dev | 09:55 | |
| *** mkollaro has quit IRC | 09:55 | |
| *** buzztroll has quit IRC | 09:56 | |
| *** saju_m has quit IRC | 09:58 | |
| *** Ryan_Lane has quit IRC | 09:59 | |
| *** marun has joined #openstack-dev | 09:59 | |
| *** jasondotstar has joined #openstack-dev | 10:00 | |
| *** yamahata has quit IRC | 10:02 | |
| *** killer_prince has quit IRC | 10:05 | |
| *** sarob has joined #openstack-dev | 10:08 | |
| *** danpb has joined #openstack-dev | 10:08 | |
| *** sarob has quit IRC | 10:14 | |
| *** rossella_s has joined #openstack-dev | 10:16 | |
| *** sarob has joined #openstack-dev | 10:18 | |
| *** bada has joined #openstack-dev | 10:18 | |
| *** sarob has quit IRC | 10:22 | |
| *** sushils has joined #openstack-dev | 10:23 | |
| *** NikitaKonovalov is now known as NikitaKonovalov_ | 10:25 | |
| *** sulrich has quit IRC | 10:28 | |
| *** sulrich has joined #openstack-dev | 10:29 | |
| *** amcrn has quit IRC | 10:29 | |
| *** paragan has quit IRC | 10:30 | |
| *** Oneiroi has quit IRC | 10:31 | |
| *** kgriffs_afk is now known as kgriffs | 10:31 | |
| *** achampion has quit IRC | 10:32 | |
| *** cnesa has quit IRC | 10:37 | |
| *** xga__ has joined #openstack-dev | 10:39 | |
| *** xga_ has quit IRC | 10:39 | |
| *** xga has quit IRC | 10:39 | |
| *** gokrokve has joined #openstack-dev | 10:39 | |
| *** xga has joined #openstack-dev | 10:40 | |
| *** kgriffs is now known as kgriffs_afk | 10:41 | |
| *** posito has joined #openstack-dev | 10:42 | |
| *** alop has quit IRC | 10:43 | |
| *** gokrokve has quit IRC | 10:44 | |
| *** oro has quit IRC | 10:46 | |
| *** jcooley_ has quit IRC | 10:52 | |
| *** jcooley_ has joined #openstack-dev | 10:52 | |
| *** cnesa has joined #openstack-dev | 10:52 | |
| *** dkuffner has quit IRC | 10:53 | |
| *** pixelb has joined #openstack-dev | 10:53 | |
| *** Adri2000 has quit IRC | 10:54 | |
| *** Adri2000 has joined #openstack-dev | 10:55 | |
| *** Adri2000 has quit IRC | 10:55 | |
| *** cnesa has quit IRC | 10:57 | |
| *** jcooley_ has quit IRC | 10:57 | |
| *** DinaBelova is now known as DinaBelova_ | 10:57 | |
| *** e0ne has joined #openstack-dev | 10:57 | |
| *** posito has quit IRC | 10:58 | |
| *** DinaBelova_ is now known as DinaBelova | 10:59 | |
| *** Adri2000 has joined #openstack-dev | 10:59 | |
| *** Adri2000 has quit IRC | 10:59 | |
| *** Adri2000 has joined #openstack-dev | 10:59 | |
| *** jasondotstar has quit IRC | 11:00 | |
| *** Adri2000 has quit IRC | 11:02 | |
| *** mrunge has quit IRC | 11:03 | |
| *** Adri2000 has joined #openstack-dev | 11:03 | |
| *** I159 has joined #openstack-dev | 11:03 | |
| *** Oneiroi has joined #openstack-dev | 11:04 | |
| *** xga__ has quit IRC | 11:04 | |
| *** xga__ has joined #openstack-dev | 11:04 | |
| *** xga has quit IRC | 11:05 | |
| *** xga has joined #openstack-dev | 11:05 | |
| *** NikitaKonovalov_ is now known as NikitaKonovalov | 11:06 | |
| *** buzztroll has joined #openstack-dev | 11:07 | |
| *** xarg has quit IRC | 11:08 | |
| *** xarg_ is now known as xarg | 11:08 | |
| *** sarob has joined #openstack-dev | 11:11 | |
| *** buzztroll has quit IRC | 11:11 | |
| *** cnesa has joined #openstack-dev | 11:14 | |
| *** rdas has joined #openstack-dev | 11:14 | |
| *** CaptTofu has joined #openstack-dev | 11:15 | |
| *** xarg_ has joined #openstack-dev | 11:15 | |
| *** sarob has quit IRC | 11:15 | |
| *** pcm_ has joined #openstack-dev | 11:15 | |
| *** pcm_ has quit IRC | 11:16 | |
| *** gcha has quit IRC | 11:17 | |
| *** CaptTofu has quit IRC | 11:19 | |
| *** pcm_ has joined #openstack-dev | 11:20 | |
| *** achampion has joined #openstack-dev | 11:21 | |
| *** pcm_ has quit IRC | 11:22 | |
| *** xga__ has quit IRC | 11:22 | |
| *** xga has quit IRC | 11:22 | |
| *** pcm_ has joined #openstack-dev | 11:23 | |
| *** romcheg is now known as romcheg_ltp | 11:24 | |
| *** romcheg_ltp is now known as romcheg | 11:24 | |
| *** jp_at_hp has joined #openstack-dev | 11:25 | |
| *** achampion has quit IRC | 11:27 | |
| *** rohitk has quit IRC | 11:28 | |
| *** boris-42_ has quit IRC | 11:30 | |
| *** kgriffs_afk is now known as kgriffs | 11:32 | |
| *** oro has joined #openstack-dev | 11:34 | |
| *** jamielennox is now known as jamielennox|away | 11:34 | |
| *** gokrokve has joined #openstack-dev | 11:39 | |
| *** mkollaro1 has quit IRC | 11:40 | |
| *** kgriffs is now known as kgriffs_afk | 11:42 | |
| *** viktors has joined #openstack-dev | 11:42 | |
| *** gokrokve has quit IRC | 11:44 | |
| *** jamespage_ has joined #openstack-dev | 11:46 | |
| *** nkinder has joined #openstack-dev | 11:49 | |
| *** colinmcnamara has joined #openstack-dev | 11:49 | |
| *** baoli has joined #openstack-dev | 11:50 | |
| *** martyntaylor has left #openstack-dev | 11:51 | |
| *** kushal has quit IRC | 11:53 | |
| *** boris-42_ has joined #openstack-dev | 11:53 | |
| *** Drankis has joined #openstack-dev | 11:53 | |
| *** jamespage_ has quit IRC | 11:55 | |
| *** baoli has quit IRC | 11:59 | |
| *** yassine has quit IRC | 12:00 | |
| *** dkuffner has joined #openstack-dev | 12:00 | |
| erlon | hi guys, we are working in a feature for swift. We are running to make it ready before the feature freeze. My question is, what is the best approach, to create the blueprint right now and then update it with the code later or, should we wait to create the blueprint when we have some code to show? | 12:03 |
| *** baoli has joined #openstack-dev | 12:06 | |
| *** baoli has quit IRC | 12:06 | |
| *** baoli has joined #openstack-dev | 12:06 | |
| *** jruzicka has joined #openstack-dev | 12:07 | |
| *** kushal has joined #openstack-dev | 12:07 | |
| *** xga has joined #openstack-dev | 12:07 | |
| *** xga__ has joined #openstack-dev | 12:07 | |
| *** yeylon__ has joined #openstack-dev | 12:08 | |
| *** cnesa has quit IRC | 12:08 | |
| *** jasondotstar has joined #openstack-dev | 12:09 | |
| *** sarob has joined #openstack-dev | 12:12 | |
| *** CaptTofu has joined #openstack-dev | 12:13 | |
| *** ytwu has quit IRC | 12:14 | |
| *** raildo has quit IRC | 12:15 | |
| *** yassine has joined #openstack-dev | 12:15 | |
| *** ytwu has joined #openstack-dev | 12:15 | |
| *** cnesa has joined #openstack-dev | 12:15 | |
| *** sarob has quit IRC | 12:17 | |
| *** sgran has quit IRC | 12:18 | |
| *** RajeshMohan has quit IRC | 12:18 | |
| *** mmagr has quit IRC | 12:19 | |
| *** tellesnobrega has quit IRC | 12:20 | |
| *** MaxV has quit IRC | 12:21 | |
| *** tellesnobrega has joined #openstack-dev | 12:21 | |
| *** bhuvan has quit IRC | 12:21 | |
| *** sarob has joined #openstack-dev | 12:22 | |
| *** ala has quit IRC | 12:22 | |
| *** aditirav has quit IRC | 12:22 | |
| *** Longgeek has joined #openstack-dev | 12:23 | |
| *** achampion has joined #openstack-dev | 12:23 | |
| *** jcooley_ has joined #openstack-dev | 12:24 | |
| *** FunnyLookinHat has joined #openstack-dev | 12:24 | |
| *** sarob has quit IRC | 12:26 | |
| *** achampion has quit IRC | 12:28 | |
| *** amotoki has quit IRC | 12:29 | |
| *** unicell has joined #openstack-dev | 12:29 | |
| *** unicell has quit IRC | 12:29 | |
| *** unicell has joined #openstack-dev | 12:29 | |
| *** bhuvan has joined #openstack-dev | 12:29 | |
| *** jcooley_ has quit IRC | 12:30 | |
| *** nkinder has quit IRC | 12:30 | |
| *** posito has joined #openstack-dev | 12:32 | |
| ihrachys | dhellmann: around? | 12:33 |
| *** kgriffs_afk is now known as kgriffs | 12:33 | |
| ihrachys | dhellmann: reading oslo meeting notes from prev Fri... so am I right that oslo devs consider oslo.messaging to be the master for 'stable' oslo-rpc implementation found in oslo-incubator? | 12:34 |
| *** jamespage_ has joined #openstack-dev | 12:35 | |
| ihrachys | dhellmann: meaning, any fix should go to oslo.messaging first and then be 'backported' to oslo-rpc | 12:35 |
| *** jamespage_ has quit IRC | 12:35 | |
| *** jasondotstar has quit IRC | 12:37 | |
| *** ctlaugh_ has joined #openstack-dev | 12:38 | |
| *** glenng has quit IRC | 12:38 | |
| *** ctlaugh has quit IRC | 12:39 | |
| *** gokrokve has joined #openstack-dev | 12:39 | |
| *** alexpilotti has joined #openstack-dev | 12:40 | |
| *** lucasagomes is now known as lucas-hungry | 12:40 | |
| *** bhuvan has quit IRC | 12:41 | |
| *** colinmcnamara has quit IRC | 12:41 | |
| *** mflobo has joined #openstack-dev | 12:42 | |
| *** fandi has quit IRC | 12:42 | |
| *** kgriffs is now known as kgriffs_afk | 12:43 | |
| *** pschaef has quit IRC | 12:43 | |
| *** gokrokve has quit IRC | 12:44 | |
| *** DinaBelova is now known as DinaBelova_ | 12:44 | |
| *** RajeshMohan has joined #openstack-dev | 12:45 | |
| *** posito has quit IRC | 12:45 | |
| *** rfolco has joined #openstack-dev | 12:46 | |
| *** nkinder has joined #openstack-dev | 12:47 | |
| *** nermina has joined #openstack-dev | 12:48 | |
| *** AnilV4 has quit IRC | 12:50 | |
| *** AnilV4 has joined #openstack-dev | 12:50 | |
| *** neeti has quit IRC | 12:52 | |
| *** IanGovett has joined #openstack-dev | 12:52 | |
| *** AMike has quit IRC | 12:54 | |
| *** rohitk has joined #openstack-dev | 12:55 | |
| *** galstrom_zzz is now known as galstrom | 12:57 | |
| *** gcha has joined #openstack-dev | 12:58 | |
| rushiagr | erlon: hey | 12:58 |
| rushiagr | erlon: you can create the blueprint before submitting the code | 12:59 |
| erlon | rushiagr: hey Rushi | 12:59 |
| *** ala has joined #openstack-dev | 13:00 | |
| *** ayoung has joined #openstack-dev | 13:00 | |
| *** bhuvan has joined #openstack-dev | 13:00 | |
| erlon | rushiagr: great | 13:01 |
| erlon | tks | 13:01 |
| rushiagr | erlon: it is also a good idea that you communicate your intent to submit code, and timeline if any, in the weekly team meeting | 13:02 |
| erlon | hmm, good, and when is this weekly meeting? | 13:02 |
| erlon | here in IRC? | 13:02 |
| *** lari_ has quit IRC | 13:02 | |
| erlon | rushiagr: ^ | 13:03 |
| *** lari_ has joined #openstack-dev | 13:03 | |
| rushiagr | erlon: it is on IRC. You can check the wiki to see the date and time of Swift meeting | 13:03 |
| rushiagr | https://wiki.openstack.org/wiki/Meetings | 13:04 |
| rushiagr | erlon: https://wiki.openstack.org/wiki/Meetings#Swift_team_meeting | 13:05 |
| *** yamahata has joined #openstack-dev | 13:05 | |
| *** nermina has quit IRC | 13:05 | |
| *** danpb has quit IRC | 13:10 | |
| *** DinaBelova_ is now known as DinaBelova | 13:11 | |
| *** david-lyle has quit IRC | 13:11 | |
| *** danpb has joined #openstack-dev | 13:12 | |
| *** nkinder has quit IRC | 13:12 | |
| *** achampion has joined #openstack-dev | 13:14 | |
| *** sarob has joined #openstack-dev | 13:15 | |
| *** bhuvan has quit IRC | 13:16 | |
| *** colinmcnamara has joined #openstack-dev | 13:16 | |
| *** athomas has joined #openstack-dev | 13:16 | |
| *** RajeshMohan has quit IRC | 13:16 | |
| *** RajeshMohan has joined #openstack-dev | 13:17 | |
| *** ytwu has quit IRC | 13:18 | |
| *** MaxV has joined #openstack-dev | 13:18 | |
| *** afazekas_ has joined #openstack-dev | 13:18 | |
| *** Drankis has quit IRC | 13:19 | |
| *** achampion has quit IRC | 13:19 | |
| *** jcooley_ has joined #openstack-dev | 13:19 | |
| *** yolanda has quit IRC | 13:19 | |
| *** MaxV has quit IRC | 13:19 | |
| *** sarob has quit IRC | 13:19 | |
| *** MaxV has joined #openstack-dev | 13:19 | |
| *** dstanek has joined #openstack-dev | 13:19 | |
| *** nermina has joined #openstack-dev | 13:22 | |
| *** markvoelker has quit IRC | 13:23 | |
| *** bhuvan has joined #openstack-dev | 13:24 | |
| *** bhuvan has quit IRC | 13:24 | |
| *** bhuvan has joined #openstack-dev | 13:24 | |
| *** jdob has joined #openstack-dev | 13:25 | |
| *** yolanda has joined #openstack-dev | 13:25 | |
| *** jcooley_ has quit IRC | 13:26 | |
| *** Clabbe has joined #openstack-dev | 13:26 | |
| *** colinmcnamara has quit IRC | 13:27 | |
| *** johnthetubaguy has joined #openstack-dev | 13:28 | |
| *** johnthetubaguy has quit IRC | 13:29 | |
| *** belmoreira has joined #openstack-dev | 13:29 | |
| *** CaptTofu has quit IRC | 13:30 | |
| *** johnthetubaguy has joined #openstack-dev | 13:30 | |
| *** cfriesen has joined #openstack-dev | 13:30 | |
| *** russellb is now known as rustlebee | 13:30 | |
| *** ndipanov is now known as ndipanoff | 13:31 | |
| *** haomaiwang has quit IRC | 13:32 | |
| *** vladikr has joined #openstack-dev | 13:33 | |
| *** haomaiwang has joined #openstack-dev | 13:33 | |
| *** afazekas_ has quit IRC | 13:33 | |
| *** kgriffs_afk is now known as kgriffs | 13:34 | |
| *** doug_shelley66 has quit IRC | 13:34 | |
| *** yassine has quit IRC | 13:34 | |
| *** sahid has quit IRC | 13:35 | |
| *** raildo has joined #openstack-dev | 13:36 | |
| *** yassine has joined #openstack-dev | 13:37 | |
| *** vartom1111111117 has quit IRC | 13:38 | |
| *** aveiga has joined #openstack-dev | 13:38 | |
| *** yeylon__ has quit IRC | 13:38 | |
| *** rohitk has quit IRC | 13:38 | |
| *** gmoro has quit IRC | 13:38 | |
| *** gokrokve has joined #openstack-dev | 13:39 | |
| *** sulrich has quit IRC | 13:39 | |
| *** milki has quit IRC | 13:39 | |
| *** alexpilotti has quit IRC | 13:40 | |
| *** milki has joined #openstack-dev | 13:43 | |
| *** gokrokve has quit IRC | 13:44 | |
| *** sahid has joined #openstack-dev | 13:44 | |
| *** kgriffs is now known as kgriffs_afk | 13:44 | |
| *** CaptTofu has joined #openstack-dev | 13:44 | |
| *** achampion has joined #openstack-dev | 13:45 | |
| *** nermina has quit IRC | 13:47 | |
| *** achampion has quit IRC | 13:47 | |
| *** jecarey has quit IRC | 13:48 | |
| *** joesavak has joined #openstack-dev | 13:49 | |
| *** jasondotstar has joined #openstack-dev | 13:50 | |
| *** pschaef has joined #openstack-dev | 13:50 | |
| *** gmoro has joined #openstack-dev | 13:50 | |
| *** rohitk has joined #openstack-dev | 13:51 | |
| *** bhuvan has quit IRC | 13:54 | |
| *** NikitaKonovalov is now known as NikitaKonovalov_ | 13:56 | |
| *** dvarga has joined #openstack-dev | 13:58 | |
| *** sulrich has joined #openstack-dev | 13:58 | |
| *** lucas-hungry is now known as lucasagomes | 13:58 | |
| *** galstrom is now known as galstrom_zzz | 13:59 | |
| *** lbragstad has quit IRC | 13:59 | |
| *** colinmcnamara has joined #openstack-dev | 14:00 | |
| *** thomasem has joined #openstack-dev | 14:00 | |
| *** bhuvan has joined #openstack-dev | 14:01 | |
| *** dbalog has joined #openstack-dev | 14:02 | |
| *** dsirrine has joined #openstack-dev | 14:02 | |
| *** gordc has joined #openstack-dev | 14:03 | |
| *** morazi has joined #openstack-dev | 14:03 | |
| *** tmclaugh[work] has joined #openstack-dev | 14:04 | |
| *** mriedem has joined #openstack-dev | 14:04 | |
| *** mfer has joined #openstack-dev | 14:04 | |
| *** alexpilotti has joined #openstack-dev | 14:05 | |
| *** jsavak has joined #openstack-dev | 14:07 | |
| *** doug_shelley66 has joined #openstack-dev | 14:09 | |
| tellesnobrega | dolphm: hi, during the hierarchical multitenancy meeting you said that having a admin domain and admin project to have a project_admin and a domain_admin role, do you think its worth to invest time in fixing this? | 14:09 |
| *** NikitaKonovalov_ is now known as NikitaKonovalov | 14:10 | |
| *** tongli has joined #openstack-dev | 14:10 | |
| *** dprince has joined #openstack-dev | 14:10 | |
| *** thuc has joined #openstack-dev | 14:10 | |
| tellesnobrega | dolphm: im trying to come up with problems that could be solved by the summit, at least with a PoC, so it could be presented there and gather opinions | 14:11 |
| *** rohitk has quit IRC | 14:11 | |
| dolphm | tellesnobrega: with hierarchical multitenancy, it might make a lot of sense to replace those rather arbitrary concepts with role assignments on the root tenant/project | 14:11 |
| *** joesavak has quit IRC | 14:11 | |
| dolphm | tellesnobrega: another approach- https://blueprints.launchpad.net/keystone/+spec/service-scoped-tokens | 14:12 |
| *** glenng has joined #openstack-dev | 14:12 | |
| *** thuc_ has joined #openstack-dev | 14:12 | |
| tellesnobrega | dolphm: i see, this role assignments are also included in vishy's PoC? | 14:12 |
| *** sgran has joined #openstack-dev | 14:12 | |
| dolphm | tellesnobrega: they are not, but it might be a logical next step | 14:13 |
| *** bknudson has quit IRC | 14:13 | |
| tellesnobrega | dolphm: good, i think i will put this on the background for now, since the concept of hierarchical multi tenancy will be defined in the summit, and im looking to bring something to the summit | 14:14 |
| *** pberis has quit IRC | 14:14 | |
| *** jcooley_ has joined #openstack-dev | 14:14 | |
| dolphm | raildo: o/ | 14:14 |
| *** igor_ has joined #openstack-dev | 14:15 | |
| *** igor__ has quit IRC | 14:15 | |
| dolphm | tellesnobrega: it's a long standing issue against keystone https://bugs.launchpad.net/keystone/+bug/968696 | 14:16 |
| *** thuc has quit IRC | 14:16 | |
| uvirtbot | Launchpad bug 968696 in keystone ""admin"-ness not properly scoped" [High,Confirmed] | 14:16 |
| *** sarob has joined #openstack-dev | 14:16 | |
| *** tdruiva has joined #openstack-dev | 14:16 | |
| *** colinmcnamara has quit IRC | 14:16 | |
| *** jaypipes has joined #openstack-dev | 14:16 | |
| raildo | dolphm: I'm working with tellesnobrega and we are trying to create a cool idea, to send to the summit = D | 14:17 |
| *** jdob has quit IRC | 14:17 | |
| *** jdob has joined #openstack-dev | 14:17 | |
| *** tdruiva has quit IRC | 14:17 | |
| dolphm | this is a topic we'll certainly be discussing - i was hoping to have a solution land in icehouse, but it's too late for that now | 14:17 |
| *** jecarey has joined #openstack-dev | 14:18 | |
| *** tdruiva has joined #openstack-dev | 14:18 | |
| *** IanGovett has quit IRC | 14:19 | |
| *** jcooley_ has quit IRC | 14:19 | |
| raildo | dolphm: Do you think a good idea to remove the requirement to be connected to domain and project to create a super admin? | 14:19 |
| *** jckasper has quit IRC | 14:20 | |
| raildo | dolphm: I think we could better discuss about it in some meeting, or create a mailing list for it | 14:20 |
| *** blues-man has joined #openstack-dev | 14:21 | |
| *** Ajaeger has joined #openstack-dev | 14:22 | |
| raildo | Send an email to the mailing list openstack-dev * | 14:22 |
| *** sarob has quit IRC | 14:22 | |
| tellesnobrega | dolphm: we are really interested in having multi tenancy on OS. Our goal is to have some improvement in this direction to show in Atlanta | 14:24 |
| *** NikitaKonovalov is now known as NikitaKonovalov_ | 14:24 | |
| *** adreznec has joined #openstack-dev | 14:25 | |
| raildo | +1 | 14:25 |
| *** sarob has joined #openstack-dev | 14:25 | |
| *** colinmcnamara has joined #openstack-dev | 14:27 | |
| *** xqueralt_ has joined #openstack-dev | 14:27 | |
| *** xqueralt_ has quit IRC | 14:28 | |
| *** xqueralt_ has joined #openstack-dev | 14:28 | |
| *** Nithya has joined #openstack-dev | 14:28 | |
| *** thuc has joined #openstack-dev | 14:29 | |
| *** sarob has quit IRC | 14:30 | |
| *** xqueralt has quit IRC | 14:31 | |
| *** xqueralt_ is now known as xqueralt | 14:31 | |
| *** lbragstad has joined #openstack-dev | 14:32 | |
| *** nshaikh has quit IRC | 14:33 | |
| *** thuc_ has quit IRC | 14:33 | |
| *** thuc has quit IRC | 14:34 | |
| *** kgriffs_afk is now known as kgriffs | 14:35 | |
| *** rdas has quit IRC | 14:35 | |
| *** florentflament has quit IRC | 14:35 | |
| *** xga__ has quit IRC | 14:36 | |
| *** xga_ has joined #openstack-dev | 14:36 | |
| *** xga has quit IRC | 14:36 | |
| *** edmund has joined #openstack-dev | 14:36 | |
| *** xga has joined #openstack-dev | 14:36 | |
| *** primeministerp has joined #openstack-dev | 14:37 | |
| *** jayg|g0n3 is now known as jayg | 14:37 | |
| *** prad_ has joined #openstack-dev | 14:38 | |
| *** bhuvan has quit IRC | 14:39 | |
| *** colinmcnamara has quit IRC | 14:39 | |
| *** doug_shelley66 has quit IRC | 14:39 | |
| *** gokrokve has joined #openstack-dev | 14:39 | |
| *** bknudson has joined #openstack-dev | 14:39 | |
| *** peristeri has joined #openstack-dev | 14:40 | |
| *** alex_xu has quit IRC | 14:40 | |
| raildo | ayoung: I was discussing with some colleagues the idea of creating a super admin for the keystone, in case it would be a user that is not related to any project or area and have the privileges to control the entire cloud. But currently, you can not do this, because every user must be associated with a project and field, for us it is conceptually wrong. So we wanted to propose the idea to the summit of being able to create a su | 14:41 |
| raildo | without being associated with any field or project and basically would include validations where there are the projects and domains, if the user has the role of super admin. You think it's valid, we spend time to propose this to the summit? | 14:41 |
| ayoung | raildo, I think it is wrong | 14:41 |
| ayoung | I think you are thinking about a different problem | 14:42 |
| *** martyntaylor has joined #openstack-dev | 14:42 | |
| ayoung | and I think you want to check out the discussion about hierarchical multitenancy on the dev mailing list | 14:42 |
| ayoung | I could see "endpoint specific roles" | 14:43 |
| *** haomaiwang has quit IRC | 14:43 | |
| ayoung | with the keystone server itself being an endpoint. As well as service level roles. | 14:43 |
| *** haomaiwang has joined #openstack-dev | 14:43 | |
| ayoung | atiwari was actually working on that, but had some misconceptions himself. | 14:43 |
| *** gokrokve has quit IRC | 14:43 | |
| ayoung | as far as the distinction between a role definition and a role assignment, but he was generally on the right track | 14:44 |
| *** athomas has quit IRC | 14:44 | |
| *** aeperezt has joined #openstack-dev | 14:44 | |
| *** kgriffs is now known as kgriffs_afk | 14:44 | |
| ayoung | raildo, everything is controlled by a rule that says "does the user have the appropriate role in this scope" | 14:44 |
| ayoung | so the question is whether the scopes we have currently defined are sufficient. I am fairly certain that they are, you just need to think about the roles themselves differently. | 14:45 |
| ayoung | However, I could see an endpoint-specific role assignment as "semantic sugar" to simplify | 14:46 |
| *** jdob_ has joined #openstack-dev | 14:46 | |
| *** jdob has quit IRC | 14:46 | |
| *** burt1 has joined #openstack-dev | 14:47 | |
| *** stan_ivanov has joined #openstack-dev | 14:47 | |
| *** rtheis has joined #openstack-dev | 14:48 | |
| *** NikitaKonovalov_ is now known as NikitaKonovalov | 14:49 | |
| *** otherwiseguy has joined #openstack-dev | 14:49 | |
| *** bhuvan has joined #openstack-dev | 14:50 | |
| *** hackeru has joined #openstack-dev | 14:50 | |
| hackeru | lol | 14:51 |
| hackeru | ? | 14:51 |
| hackeru | ?? | 14:51 |
| *** hackeru has left #openstack-dev | 14:51 | |
| dolphm | ayoung: if you consider services to be inherent owners of their own data, then it makes sense for all tenants/projects to inherit role assignments from a root tenant which owns the services themselves | 14:52 |
| *** hackeru has joined #openstack-dev | 14:52 | |
| tellesnobrega | ayoung: what raildo is suggesting it to be able to have a user that won't be attached to any domains or projects, and will have a super_admin role, and this role will be inherited in all domains and projects, so he will be able to manage the whole cloud. He will need to get a token to whichever scope he wants to manage | 14:52 |
| dolphm | i'm sure deployments like HP's will happily complicate things further by having hierarchies of services | 14:52 |
| ayoung | tellesnobrega, we've been there before | 14:52 |
| ayoung | it wa called global roles. | 14:52 |
| raildo | I see today is that the roles are not sufficient to elaborate a better cloud, then the implementation of the new policy.json v3, where we can create a domain admin or project admin, for example, will greatly improve it but when we treat the super admin, I think this is kinda weird being required to authenticate a project to list the roles, since it will not actually run anything on any project | 14:52 |
| raildo | But I'll read the best and suggested discussion about the endpoint-specific role assign, to see how I can help with that. | 14:53 |
| ayoung | Better, and more consistent, to have a container for a role | 14:53 |
| *** dperaza1 has joined #openstack-dev | 14:53 | |
| *** jruzicka is now known as limited_cocotte | 14:53 | |
| tellesnobrega | ayoung: i see | 14:53 |
| dolphm | raildo: the concept of global role assignments contradicts the rules of multitenancy | 14:53 |
| ayoung | Bottom line is, yes, there needs to be something at the root of the tree | 14:54 |
| hackeru | =)))) | 14:54 |
| *** FunnyLookinHat has quit IRC | 14:55 | |
| *** athomas has joined #openstack-dev | 14:55 | |
| *** dperaza has quit IRC | 14:55 | |
| *** limited_cocotte is now known as transcendence | 14:56 | |
| *** dansmith is now known as damnsmith | 14:56 | |
| *** transcendence is now known as jruzicka | 14:56 | |
| *** dims has quit IRC | 14:57 | |
| tellesnobrega | ayoung and dolphm and raildo, from what im getting from this, and i may be wrong (correct if i am), working with roles right now isn't a good plan, since the definition on hierarchical multitenancy will define how roles will work from this point on. | 14:57 |
| raildo | dolphm: I understand. Do you have any idea how it would be a super admin using multitenancy? In the context of hierarchical multitenancy , super admin would be associated with the project and have root access to their children, grandchildren ... to the leaves, and project_admin would have access only to your project, right? | 14:57 |
| ayoung | so... domain is the root of the tree for multitenancy | 14:57 |
| dolphm | tellesnobrega: ++ | 14:57 |
| ayoung | think of a domain as a project without a parent | 14:57 |
| *** Ajaeger has left #openstack-dev | 14:57 | |
| dolphm | ayoung: -- | 14:58 |
| ayoung | feh | 14:58 |
| dolphm | close, but... | 14:58 |
| *** byeager has joined #openstack-dev | 14:58 | |
| ayoung | dolphm, it is the top level namespace as well. | 14:58 |
| dolphm | ayoung: as discussed in last week's meeting, all current domains would basically inherit from a single root/null node | 14:58 |
| dolphm | ayoung: so you can only have a single hierarchy | 14:58 |
| *** oro has quit IRC | 14:58 | |
| ayoung | And it gets dropped off the namespace for a project so that tenantA doesnt see too much info about tenantB | 14:58 |
| *** markvoelker has joined #openstack-dev | 14:59 | |
| raildo | What need a domain? I think the valid existence because the features that exist today but conceptually this does not seem right. | 14:59 |
| ayoung | raildo, namespacing | 14:59 |
| *** tsekiyama has joined #openstack-dev | 15:00 | |
| ayoung | coke and pepsi in the same datacenter | 15:00 |
| ayoung | each get a project called "general" | 15:00 |
| ayoung | prior to domain, was not possible | 15:00 |
| ayoung | it would have to be flat, like variable names in C | 15:00 |
| *** kevinconway has joined #openstack-dev | 15:00 | |
| ayoung | coke_general and pepsi_general | 15:00 |
| *** jmckind has joined #openstack-dev | 15:00 | |
| *** Gordonz has joined #openstack-dev | 15:01 | |
| *** n0ano has quit IRC | 15:01 | |
| mtreinish | markmcclain: hey, I'm around now | 15:02 |
| *** IanGovett has joined #openstack-dev | 15:03 | |
| *** vijendar has joined #openstack-dev | 15:03 | |
| *** eharney has joined #openstack-dev | 15:04 | |
| raildo | ayoung: e dolphm thanks for the help, we will talk here, think of something useful and returned to discuss with you (if it is not bothering you too much) | 15:04 |
| ayoung | Not at all | 15:04 |
| dolphm | raildo: happy to talk through it | 15:04 |
| tellesnobrega | tellesnobrega: appreciate the help as well | 15:05 |
| *** doug_shelley66 has joined #openstack-dev | 15:06 | |
| *** tkammer has joined #openstack-dev | 15:06 | |
| mtreinish | markmcclain: hey, I'm around now | 15:07 |
| *** ArxCruz has quit IRC | 15:07 | |
| *** iartarisi has quit IRC | 15:07 | |
| *** zul has quit IRC | 15:07 | |
| *** vijendar has quit IRC | 15:08 | |
| *** jcooley_ has joined #openstack-dev | 15:08 | |
| YorikSar | Hello, everyone. I want to raise here a question I've already asked on ML: Do we really need eventlet (or asyncio or whatever) in our projects? http://lists.openstack.org/pipermail/openstack-dev/2014-February/026568.html | 15:08 |
| *** stevemar has joined #openstack-dev | 15:08 | |
| *** rbrady is now known as rbrady-mtg | 15:08 | |
| YorikSar | A little followup here: http://lists.openstack.org/pipermail/openstack-dev/2014-February/026583.html | 15:09 |
| *** ArxCruz has joined #openstack-dev | 15:09 | |
| *** zaitcev has joined #openstack-dev | 15:10 | |
| *** zul has joined #openstack-dev | 15:10 | |
| *** dims has joined #openstack-dev | 15:10 | |
| YorikSar | ayoung: I'm really interested in your opinion on that matter as Keystone have been shifted away from eventlet thanks to you. | 15:10 |
| *** boris-42_ has quit IRC | 15:12 | |
| *** bvandenh has quit IRC | 15:13 | |
| ayoung | YorikSar, no good deed goes unpunished. I've been telling people to move to HTTPD for years (2 to be exact) and now that they are, we find mod_wsgi has a hardcoded header size limit that prevents the tokens from holding large catalogs | 15:13 |
| *** boris-42_ has joined #openstack-dev | 15:13 | |
| ayoung | so...I'm wokring on compressing tokens | 15:13 |
| ayoung | eventlet and continuation based web servers make sense for certain workloads, but read/write DB apps are not one of them | 15:14 |
| ayoung | with Keystone tokens going crypto, we incurr a higher CPU load, and greenthreads don't have a good answer for that either | 15:14 |
| *** jcooley_ has quit IRC | 15:14 | |
| *** ijw has joined #openstack-dev | 15:14 | |
| *** schwicke has joined #openstack-dev | 15:14 | |
| ayoung | async io is a tool in the toolbox, but as a clever man once observed "There is no silver bullet" in programming. | 15:14 |
| YorikSar | ayoung: I think it's kept around as just some tool with no actual use. | 15:16 |
| *** ogelbukh has quit IRC | 15:16 | |
| ayoung | interia is a powerful force | 15:16 |
| YorikSar | ayoung: Btw, why not deploy Keystone with some other server then? | 15:17 |
| ayoung | YorikSar, other than HTTPD? | 15:17 |
| YorikSar | ayoung: Yes | 15:17 |
| ayoung | its a bug in mod_wsgi | 15:17 |
| ayoung | packaging | 15:17 |
| hackeru | "___" | 15:17 |
| YorikSar | ayoung: Oh... | 15:17 |
| ayoung | YorikSar, a better approach is to replace mod_wsgi, I think | 15:17 |
| *** kragniz has quit IRC | 15:18 | |
| ayoung | there is an alternative... | 15:18 |
| hackeru | ;) | 15:18 |
| YorikSar | ayoung: You mean, it can be repackaged so that we'll get bigger header limit? | 15:18 |
| ayoung | heh, I meant that switching web servers is a packaing headache for the major distros | 15:18 |
| ayoung | easier to patch mod_wsgi | 15:18 |
| *** mkollaro has joined #openstack-dev | 15:19 | |
| *** sweston has joined #openstack-dev | 15:19 | |
| ayoung | I can't remember the name, but there is a different module...looking | 15:19 |
| YorikSar | ayoung: There is mod_python, but it's an old dead thing. | 15:20 |
| ayoung | nah, something new and shiny that I don';t yet trust | 15:20 |
| ayoung | dolphm, what was the alternative to mod_wsgi people were looking for support on? | 15:20 |
| *** mikeoutland has joined #openstack-dev | 15:20 | |
| ayoung | gunicorn | 15:21 |
| dolphm | ayoung: chaussette? | 15:21 |
| *** dvarga is now known as dvarga|away | 15:21 | |
| dolphm | ayoung: i don't know -- *any* other wsgi server? | 15:21 |
| *** dvarga|away is now known as dvarga | 15:21 | |
| ayoung | that is a new one to me dolphm | 15:21 |
| ayoung | I think it is gunicorn | 15:21 |
| dolphm | ayoung: they're all the same | 15:21 |
| ayoung | ah | 15:21 |
| dolphm | in terms of how we support them | 15:21 |
| *** ijw has quit IRC | 15:22 | |
| YorikSar | Ok, so they're basically separate servers talking to Apache frontend (if it's needed). | 15:22 |
| YorikSar | uWSGI is a cool one, for example. | 15:22 |
| ayoung | YorikSar, yeah | 15:22 |
| ayoung | but prefork | 15:22 |
| *** ijw has joined #openstack-dev | 15:22 | |
| YorikSar | ayoung: Some of them (gunicorn, for example) can for, spawn threads and run eventlet hubs in them. | 15:23 |
| *** colinmcnamara has joined #openstack-dev | 15:23 | |
| YorikSar | *fork | 15:23 |
| ayoung | so long as the SSL and other crypto is done native, prior to python code, I'm OK with it (I think) | 15:23 |
| *** andreykurilin has joined #openstack-dev | 15:23 | |
| ayoung | I want SSL everywhere and strong crypto authentication | 15:23 |
| YorikSar | ayoung: Yeah, encrypt and authenticate everything - that's why you're in Keystone, right? :) | 15:24 |
| *** mikeoutland has quit IRC | 15:25 | |
| YorikSar | But... Getting back to my question. | 15:25 |
| YorikSar | Do you see any use of async in core projects? | 15:25 |
| *** browne has joined #openstack-dev | 15:26 | |
| *** rraja has joined #openstack-dev | 15:26 | |
| *** david-lyle has joined #openstack-dev | 15:26 | |
| YorikSar | Async is cool when there's a lot (really lot) network connections with really lot data transfer. | 15:26 |
| YorikSar | I don't see that happening in OpenStack. | 15:26 |
| therve | Really? | 15:27 |
| YorikSar | therve: What do you have in mind? | 15:28 |
| *** xqueralt has quit IRC | 15:29 | |
| therve | OpenStack is all about interacting APIs, if that's not a lot of network connections I don't know what that is | 15:29 |
| *** xqueralt has joined #openstack-dev | 15:29 | |
| *** Nikolay_St has joined #openstack-dev | 15:32 | |
| *** ogelbukh has joined #openstack-dev | 15:32 | |
| YorikSar | therve: A lot is P2P full-mesh connections in huge DHT network. OpenStack API servers won't handle 10k simultanious requests at a time because a) requests are handled too quickly and b) there's actually no need for it. | 15:33 |
| *** florentflament has joined #openstack-dev | 15:33 | |
| ayoung | YorikSar, so, it a task cannot be completed immediately, it should be recorded and the user should get back a 202 | 15:33 |
| ayoung | now...question is what is meant by "recorded" | 15:33 |
| *** galstrom_zzz is now known as galstrom | 15:34 | |
| ayoung | I probably it means appended to a high throughput log | 15:34 |
| *** jgrimm has joined #openstack-dev | 15:34 | |
| YorikSar | ayoung: Services usually send an RPC request to some backend worker. | 15:34 |
| ayoung | you can put it in a database, so long as it can be injected without conflicts | 15:34 |
| ayoung | RPC is problematic | 15:34 |
| therve | YorikSar, I don't understand your point. It seems to be "We shouldn't care about scaling in OpenStack", which seems weird to me | 15:34 |
| ayoung | you don;t want to wait for a remote call | 15:34 |
| YorikSar | ayoung: Or store something in DB (which won't yield greenthread btw) | 15:35 |
| ayoung | therve, we should not worry about scaling in openstack | 15:35 |
| ayoung | right | 15:35 |
| *** mrodden1 is now known as mrodden | 15:35 | |
| ayoung | therve, we should worry about scaling across openstack | 15:35 |
| *** athomas has quit IRC | 15:35 | |
| ayoung | but scalability can't come with a sacrifice of security or stability | 15:35 |
| *** kgriffs_afk is now known as kgriffs | 15:35 | |
| therve | I don't understand the difference | 15:36 |
| ayoung | therve, good, I've given you something to thinkabout then | 15:36 |
| YorikSar | therve: No. I suggest not to worry about scaling just because. I suggest to remove eventlet from and let Apache HTTPD, gunicorn, uWSGI, or whichever other application server handle scaling in production. | 15:36 |
| *** hackeru has left #openstack-dev | 15:36 | |
| ayoung | YorikSar, if we have to wait for an RPC, async IO off system, or even a write to local disk that is going to have contention, the request might as well be handled by a separate thread | 15:37 |
| ayoung | the OS is going to be best capable of handling that | 15:37 |
| therve | YorikSar, Those are not application servers, but web servers. | 15:37 |
| YorikSar | ayoung: Actually if you're waiting for RPC call, you thread will yield (GIL released, or eventlet's hub take it over). | 15:37 |
| *** JoshuaG_AIM has left #openstack-dev | 15:38 | |
| *** Nikolay_St has quit IRC | 15:38 | |
| ayoung | YorikSar, yes, assuming that all of the underlying code is greenlet friendly | 15:38 |
| YorikSar | therve: It depends on definition | 15:38 |
| ayoung | but as soon as you need to call crypto, all bets are off | 15:38 |
| ayoung | and, for RPC calls, you need to call crypto, or you get no message signing, | 15:38 |
| ayoung | I hate the term RPC | 15:39 |
| ayoung | lets call it posting to a queue | 15:39 |
| *** achampion has joined #openstack-dev | 15:39 | |
| *** gokrokve has joined #openstack-dev | 15:39 | |
| *** pablosan has quit IRC | 15:39 | |
| *** alop has joined #openstack-dev | 15:39 | |
| YorikSar | ayoung: That's exactly my point: stop worrying about eventlet already. It's at least not helping when we cannot guarantee that all libs are eventlet-friendly. | 15:39 |
| ayoung | when you post a message to a queue from a web server, you want to sign the message first. that is what the KDS work is all about. Signing a message requires callling in to crypto library, the Bets" you can get from eventlet is to do a process fork and wait | 15:40 |
| dstanek | YorikSar: you have the same problem with gunicorn and others | 15:40 |
| *** pablosan has joined #openstack-dev | 15:40 | |
| ayoung | but, even that, it turns out is problematic | 15:40 |
| *** jergerber has joined #openstack-dev | 15:40 | |
| ayoung | dstanek, not if you have already forked | 15:40 |
| YorikSar | ayoung: And with asyncio things become even worse since even less libs are going to be asyncio-compatible. | 15:40 |
| *** FunnyLookinHat has joined #openstack-dev | 15:40 | |
| *** jprovazn is now known as jprovazn_bbl | 15:41 | |
| dstanek | ayoung: foking has nothing to do with libs being greenlet friendly | 15:41 |
| YorikSar | ayoung: You can have threads for this, actually. | 15:41 |
| ayoung | with a prefork model, you can just do in process crypto, and who cares about the GIL etc at that point | 15:41 |
| ayoung | dstanek, I think you are missing the point | 15:41 |
| ayoung | noin-greenlet friendly libs are usable in a prefork model | 15:42 |
| YorikSar | ayoung: crypto libs should release GIL on CPU-intensive operations. | 15:42 |
| ayoung | YorikSar, then you need to have Python specific wrappers for everything you call. | 15:42 |
| *** jnoller has joined #openstack-dev | 15:42 | |
| YorikSar | ayoung: What do you mean? | 15:42 |
| *** vijendar has joined #openstack-dev | 15:43 | |
| *** colinmcnamara has quit IRC | 15:43 | |
| ayoung | YorikSar, releaseing the GIL must be done in native code. | 15:43 |
| *** mikeoutland has joined #openstack-dev | 15:43 | |
| YorikSar | ayoung: How do you call a library without a wrapper? | 15:43 |
| danpb | ayoung: or wrap all the native API calls in eventlet native threadpool as we do for libvirt | 15:43 |
| ayoung | which means that a general purpose library then needs a python specific wrapper | 15:43 |
| ayoung | which is fine, if it exists | 15:43 |
| YorikSar | ayoung: Do you have some specific crypto library that is GIL-greedy? | 15:44 |
| *** gokrokve has quit IRC | 15:44 | |
| ayoung | danpb, if the majority of the work is going to require a threadpool, why even bother with eventlet? | 15:44 |
| *** tjones has joined #openstack-dev | 15:44 | |
| *** alop has quit IRC | 15:44 | |
| ayoung | YorikSar, I don;t want to have to care | 15:44 |
| *** mikeoutland has quit IRC | 15:45 | |
| danpb | ayoung: well that is a good point - i frequently wish we use real threads everywhere instead ofo eventlet, but ho hum that decision was made a while back | 15:45 |
| ayoung | YorikSar, people used to complain about Java that it reimplemented everything in Java. I don't want to have the same complaint about Python | 15:45 |
| *** kgriffs is now known as kgriffs_afk | 15:45 | |
| YorikSar | ayoung: Well... If you use some library with native calls, it'll most likely release GIL on CPU-intensive operations that doesn't touch Python objects, IO operations, etc. | 15:45 |
| YorikSar | ayoung: If it doesn't, it's already considered a bad library/wrapper. | 15:46 |
| ayoung | danpb, So We've worked hard in Keystone to make it such that it can be run either in an eventlet or non-eventlet based WSGI container, and that at startup time, you have an explicit call to determine which model you are using | 15:46 |
| *** alop has joined #openstack-dev | 15:46 | |
| *** vijendar has quit IRC | 15:47 | |
| *** tkammer has quit IRC | 15:47 | |
| YorikSar | But I'm not actually advocating for threaded model. | 15:47 |
| ayoung | YorikSar, the point is, for Keystone, HTTPD in prefork is logical. If we need to then scale out to threads, it requires an additional level of complexity analysis. Adding in greenthreads takes it yet again to another level. | 15:47 |
| *** vijendar has joined #openstack-dev | 15:47 | |
| *** kolesovdv has quit IRC | 15:48 | |
| YorikSar | Although it beats forking model in memory consumption. | 15:48 |
| ayoung | http://c2.com/cgi/wiki?PrematureOptimization | 15:48 |
| *** galstrom is now known as galstrom_zzz | 15:48 | |
| YorikSar | ayoung: What do you mean by complexity analysis? Like benchmarking or like looking for races? | 15:49 |
| *** colinmcnamara has joined #openstack-dev | 15:50 | |
| YorikSar | ayoung: And yes, I think that using threads, especially the green ones is a one huge premature optimization that has been done extremely early in OpenStack life. | 15:50 |
| ayoung | ++ | 15:50 |
| *** yaguang has quit IRC | 15:51 | |
| YorikSar | ayoung: That's why I really suggest to move away from any async stuff in our code instead of spending time replacing eventlet with whatever. | 15:51 |
| *** rraja_ has joined #openstack-dev | 15:51 | |
| *** vijendar has quit IRC | 15:52 | |
| *** vijendar has joined #openstack-dev | 15:52 | |
| *** aveiga has quit IRC | 15:53 | |
| *** rraja has quit IRC | 15:53 | |
| ayoung | YorikSar, there was a battle about 10 years ago for the soul of the threading model in Linux. On one side was the "thread in kernel space" people, and on the other "thread in user space" and the Kernel folks won out. So we have people that think that we need to thread in userspace without Kernel support. I'm of the opinion that we need to think in terms of web servers that can be massively scaled horizontally. Lets not opti | 15:53 |
| ayoung | mize the servers for performance, lets optimize them for statelessness. | 15:53 |
| *** kushal has quit IRC | 15:53 | |
| dstanek | ayoung: in most production deployments i have used gunicorn to prefork based on # of CPUs - with gunicorn the workers run with gevent | 15:53 |
| ayoung | that is "cloud" | 15:53 |
| *** pablosan_ has joined #openstack-dev | 15:54 | |
| *** pablosan has quit IRC | 15:54 | |
| *** VINOD_ has joined #openstack-dev | 15:54 | |
| *** carl_baldwin has joined #openstack-dev | 15:54 | |
| dstanek | ayoung: ++ | 15:54 |
| ayoung | dstanek, I've been so "all over the place" in my career that I can't say what I've done in "most" deployments. | 15:54 |
| YorikSar | ayoung: It looks like we're in the same camp on this topic. If you have some time could you please chime in on my thread in the ML? You word would give more weight to my point. | 15:55 |
| ayoung | OK...back to my day job | 15:55 |
| *** atiwari has joined #openstack-dev | 15:55 | |
| ayoung | YorikSar, I don;t think it matters at the "openstack" level. I think each individual project has to commit to being able to run on servers other than eventlet or it is not going to happen | 15:55 |
| ayoung | I can barely cover the Keystone cases | 15:55 |
| *** amerine_ has quit IRC | 15:55 | |
| dstanek | ayoung: i've been lucky (you might say unlucky) enough to have work almost exclusively in Python for the last 12 years | 15:56 |
| ayoung | people on swift are pushing there | 15:56 |
| ayoung | dstanek, python is a decent subset of Lisp, but it lacks a macro preprocessor | 15:56 |
| *** mikeoutland has joined #openstack-dev | 15:56 | |
| ayoung | :) | 15:56 |
| dstanek | :) | 15:56 |
| *** tmclaugh[work] has quit IRC | 15:57 | |
| YorikSar | ayoung: Well... I'm not sure that it doesn't matter on the OpenStack level. | 15:57 |
| ayoung | YorikSar, are you heads down lookking at Nova code? THat is the place that I care about it the most, and I have no time to get dirty with it, | 15:58 |
| YorikSar | ayoung: Because if people invest a lot of time into shifting to asyncio, it'll be even harder to convince them to move away from it. | 15:58 |
| *** dencaval has joined #openstack-dev | 15:58 | |
| *** jmontemayor has joined #openstack-dev | 15:58 | |
| ayoung | YorikSar, actually, the consumers of the Messaging code are problematic. Those use greenthreads, and probably need to move to Posix threads. I am not as concerned about the Web APIS | 15:59 |
| ayoung | its the pure message driven code that should really be posix threaded | 15:59 |
| YorikSar | I already had one PTL telling me "Why the hell do we need this? All other projects use eventlet. It works for us as well. An hour spent on this topic is a waste of time." | 15:59 |
| ayoung | YorikSar, people can propose, but it still needs to get by code review | 15:59 |
| ayoung | YorikSar, fix Nova, and the rest will follow | 16:00 |
| *** zul has quit IRC | 16:00 | |
| YorikSar | ayoung: I don't understand, what is your consern about messaging? | 16:00 |
| *** Ruetobas has quit IRC | 16:01 | |
| YorikSar | ayoung: It can run inside application server just as WSGI app would. | 16:01 |
| *** zul has joined #openstack-dev | 16:01 | |
| *** troytoman-away is now known as troytoman | 16:01 | |
| *** jcooley_ has joined #openstack-dev | 16:01 | |
| ayoung | YorikSar, bigger topic than I have time for now. I have an internal meeting in a few minutes | 16:01 |
| *** tonix has joined #openstack-dev | 16:02 | |
| *** athomas has joined #openstack-dev | 16:02 | |
| *** raildo has quit IRC | 16:02 | |
| *** raildo has joined #openstack-dev | 16:02 | |
| YorikSar | ayoung: For example, uWSGI can be configured to run as much processes/threads as you want and quickly spawn new ones without having to go back to HTTP. | 16:02 |
| *** tkammer has joined #openstack-dev | 16:02 | |
| YorikSar | ayoung: Oh, sure. Thanks for your time. I'm glad that I'm not alone with this mindset. | 16:02 |
| *** Ruetobas has joined #openstack-dev | 16:03 | |
| *** pmathews has joined #openstack-dev | 16:03 | |
| *** amerine has joined #openstack-dev | 16:03 | |
| *** tjones has quit IRC | 16:04 | |
| *** xarses has quit IRC | 16:04 | |
| *** xga_ has quit IRC | 16:04 | |
| *** mflobo has quit IRC | 16:05 | |
| *** tjones has joined #openstack-dev | 16:05 | |
| *** pradeep has left #openstack-dev | 16:07 | |
| *** kushal has joined #openstack-dev | 16:07 | |
| *** jcooley_ has quit IRC | 16:08 | |
| *** Ruetobas has quit IRC | 16:08 | |
| *** david-lyle has quit IRC | 16:08 | |
| *** mikeoutland has quit IRC | 16:09 | |
| dstanek | ayoung: don't you love these really big and complicated reviews..https://review.openstack.org/71932 | 16:10 |
| *** pablosan_ has quit IRC | 16:11 | |
| *** pablosan has joined #openstack-dev | 16:11 | |
| *** dkuffner has quit IRC | 16:11 | |
| *** tjones has quit IRC | 16:12 | |
| ayoung | dstanek, I think you are just submitting that to get the ATC discount at the summit | 16:13 |
| *** troytoman is now known as troytoman-away | 16:13 | |
| *** Ruetobas has joined #openstack-dev | 16:13 | |
| *** capri has quit IRC | 16:13 | |
| dstanek | ayoung: ssshhhh | 16:14 |
| *** AlexF has joined #openstack-dev | 16:14 | |
| *** capri has joined #openstack-dev | 16:14 | |
| *** tkammer has quit IRC | 16:17 | |
| *** thouveng has quit IRC | 16:17 | |
| dhellmann | ihrachys: yes, you have that right about oslo.messaging | 16:17 |
| *** salv-orlando has quit IRC | 16:17 | |
| *** athomas has quit IRC | 16:17 | |
| *** tmclaugh[work] has joined #openstack-dev | 16:17 | |
| *** mikeoutland has joined #openstack-dev | 16:18 | |
| *** cnesa has quit IRC | 16:18 | |
| *** salv-orlando has joined #openstack-dev | 16:18 | |
| *** afazekas_ has joined #openstack-dev | 16:18 | |
| *** adnan has joined #openstack-dev | 16:19 | |
| *** jamieh has quit IRC | 16:19 | |
| *** jamieh has joined #openstack-dev | 16:21 | |
| *** jobewan has joined #openstack-dev | 16:22 | |
| *** xqueralt has quit IRC | 16:23 | |
| *** bnemec is now known as beekneemech | 16:23 | |
| *** bswartz has quit IRC | 16:24 | |
| *** armax has joined #openstack-dev | 16:26 | |
| *** bauzas has quit IRC | 16:26 | |
| *** bdpayne has joined #openstack-dev | 16:29 | |
| *** romcheg1 has joined #openstack-dev | 16:30 | |
| *** romcheg1 is now known as romcheg_ltp | 16:30 | |
| *** rraja_ has quit IRC | 16:31 | |
| *** mlavalle has joined #openstack-dev | 16:32 | |
| *** cdub has joined #openstack-dev | 16:34 | |
| *** PaulMurray has quit IRC | 16:34 | |
| *** ijw has quit IRC | 16:34 | |
| *** devoid has joined #openstack-dev | 16:35 | |
| *** morazi has quit IRC | 16:36 | |
| *** kgriffs_afk is now known as kgriffs | 16:36 | |
| *** ijw has joined #openstack-dev | 16:36 | |
| *** lcheng_ has joined #openstack-dev | 16:37 | |
| *** buzztroll has joined #openstack-dev | 16:37 | |
| *** thuc has joined #openstack-dev | 16:38 | |
| *** gokrokve has joined #openstack-dev | 16:39 | |
| *** ppetit has quit IRC | 16:40 | |
| *** morazi has joined #openstack-dev | 16:40 | |
| *** viktors has left #openstack-dev | 16:40 | |
| *** ijw has quit IRC | 16:41 | |
| *** sballe has joined #openstack-dev | 16:41 | |
| *** mjfork has quit IRC | 16:42 | |
| *** rraja_ has joined #openstack-dev | 16:42 | |
| *** CaptTofu has quit IRC | 16:43 | |
| *** YorikSar has quit IRC | 16:43 | |
| *** pasquier-s has quit IRC | 16:43 | |
| *** CaptTofu has joined #openstack-dev | 16:43 | |
| *** pasquier-s_ has quit IRC | 16:43 | |
| *** gokrokve has quit IRC | 16:44 | |
| *** aaronjamesford has joined #openstack-dev | 16:45 | |
| *** xmltok has joined #openstack-dev | 16:46 | |
| *** kgriffs is now known as kgriffs_afk | 16:46 | |
| *** SumitNaiksatam has quit IRC | 16:47 | |
| *** nati_ueno has joined #openstack-dev | 16:47 | |
| *** CaptTofu has quit IRC | 16:48 | |
| *** terrylhowe has joined #openstack-dev | 16:48 | |
| *** e0ne has quit IRC | 16:48 | |
| *** CaptTofu has joined #openstack-dev | 16:49 | |
| *** markmcclain has joined #openstack-dev | 16:49 | |
| *** sushils has quit IRC | 16:49 | |
| *** sahid has quit IRC | 16:51 | |
| *** mikeoutland has quit IRC | 16:52 | |
| *** tjones has joined #openstack-dev | 16:52 | |
| *** taps has joined #openstack-dev | 16:52 | |
| *** xqueralt has joined #openstack-dev | 16:53 | |
| *** smurugesan has joined #openstack-dev | 16:53 | |
| *** lcheng_ has quit IRC | 16:54 | |
| *** nelsnelson has joined #openstack-dev | 16:55 | |
| *** _cjones_ has joined #openstack-dev | 16:55 | |
| *** mjfork has joined #openstack-dev | 16:55 | |
| *** ijw has joined #openstack-dev | 16:55 | |
| *** marcoemorais has joined #openstack-dev | 16:56 | |
| *** armax has left #openstack-dev | 16:56 | |
| *** cnesa has joined #openstack-dev | 16:57 | |
| *** VINOD_ has quit IRC | 16:57 | |
| *** afazekas_ has quit IRC | 16:57 | |
| *** schwicke has quit IRC | 16:58 | |
| *** rraja_ has quit IRC | 16:58 | |
| *** VINOD has joined #openstack-dev | 16:58 | |
| *** tiamar has quit IRC | 16:58 | |
| *** VINOD has quit IRC | 16:58 | |
| *** csd has joined #openstack-dev | 16:58 | |
| *** jcooley_ has joined #openstack-dev | 16:59 | |
| *** pschaef has quit IRC | 16:59 | |
| *** doug-fish has quit IRC | 17:00 | |
| *** markmc has quit IRC | 17:01 | |
| *** gokrokve has joined #openstack-dev | 17:01 | |
| *** kbrierly has quit IRC | 17:01 | |
| *** obondarev has quit IRC | 17:02 | |
| *** kbrierly has joined #openstack-dev | 17:02 | |
| *** obondarev has joined #openstack-dev | 17:03 | |
| *** ijw has quit IRC | 17:03 | |
| *** jcooley_ has quit IRC | 17:03 | |
| *** NikitaKonovalov is now known as NikitaKonovalov_ | 17:03 | |
| stevemar | bknudson, ping | 17:04 |
| bknudson | stevemar: what's up? | 17:04 |
| *** yassine has quit IRC | 17:04 | |
| *** ijw has joined #openstack-dev | 17:04 | |
| stevemar | bknudson, not sure if you had a chance to read my rambling comments on the ruleprocessor patch, wanted to chat about how to improve it | 17:04 |
| *** mkerrin has quit IRC | 17:05 | |
| *** doug-fish has joined #openstack-dev | 17:05 | |
| *** vuil has joined #openstack-dev | 17:05 | |
| jnoller | anyone got a link for a project's blueprint that are *good* - there's a lot of ones that seem light on information or super heavyweight | 17:05 |
| *** xarses has joined #openstack-dev | 17:05 | |
| *** vuil has quit IRC | 17:05 | |
| *** gokrokve has quit IRC | 17:06 | |
| bknudson | stevemar: ok, so maybe shouldn't return if "any_one_of", but continue to the next requirement | 17:06 |
| *** cnesa has quit IRC | 17:06 | |
| bknudson | using continue | 17:07 |
| *** rgerganov has quit IRC | 17:07 | |
| stevemar | bknudson, right, I was thinking about that last night, but returning content is still an issue | 17:08 |
| *** kgriffs_afk is now known as kgriffs | 17:08 | |
| stevemar | bknudson, I was thinking, maybe do the local variable replacement, if it's neither any_one_of or not_any_of, and just return true/false for that function | 17:08 |
| *** Mandell has quit IRC | 17:08 | |
| *** mrodden has quit IRC | 17:08 | |
| *** cnesa has joined #openstack-dev | 17:09 | |
| *** DinaBelova is now known as DinaBelova_ | 17:09 | |
| *** belmoreira has quit IRC | 17:09 | |
| bknudson | stevemar: it would be nice to only have to look at them once. | 17:09 |
| *** feleouet has quit IRC | 17:09 | |
| *** jasondotstar has quit IRC | 17:09 | |
| stevemar | agreed | 17:09 |
| *** vuil has joined #openstack-dev | 17:10 | |
| stevemar | bknudson, but that would mean overwriting what comes in? | 17:10 |
| bknudson | stevemar: it does "direct_maps += direct_map_value" -- couldn't it do "direct_maps += do_replacement(direct_map_value)" ? | 17:11 |
| stevemar | bknudson, yeah, thats what i'm getting what | 17:12 |
| *** david-lyle has joined #openstack-dev | 17:12 | |
| stevemar | bknudson, but you mentioned you didn't like over-writing the values of the input data | 17:12 |
| bknudson | stevemar: direct_maps is a new list -- direct_maps = [] | 17:13 |
| bknudson | adding to it isn't going to overwrite anything. | 17:13 |
| *** tanisdl has joined #openstack-dev | 17:13 | |
| *** kbrierly has quit IRC | 17:13 | |
| *** kbrierly has joined #openstack-dev | 17:14 | |
| *** amerine has quit IRC | 17:15 | |
| *** safchain has quit IRC | 17:15 | |
| *** Gordonz has quit IRC | 17:16 | |
| *** david-lyle has quit IRC | 17:17 | |
| *** martyntaylor has left #openstack-dev | 17:17 | |
| *** zzelle has quit IRC | 17:18 | |
| *** ygbo has quit IRC | 17:18 | |
| *** AlexF has quit IRC | 17:20 | |
| *** matrohon has quit IRC | 17:20 | |
| *** xgsa has quit IRC | 17:20 | |
| *** jpich has quit IRC | 17:20 | |
| *** dshulyak has quit IRC | 17:20 | |
| *** xga has quit IRC | 17:21 | |
| *** MaxV has quit IRC | 17:21 | |
| *** ndipanoff has quit IRC | 17:22 | |
| *** pablosan has quit IRC | 17:22 | |
| *** chris_johnson has joined #openstack-dev | 17:23 | |
| *** godara has joined #openstack-dev | 17:24 | |
| *** gcha has quit IRC | 17:24 | |
| *** rbrady-mtg is now known as rbrady | 17:24 | |
| *** martyntaylor has joined #openstack-dev | 17:24 | |
| *** mjfork_ has joined #openstack-dev | 17:24 | |
| *** mrodden has joined #openstack-dev | 17:24 | |
| *** hemnafk is now known as hemna | 17:25 | |
| *** pvaneck has joined #openstack-dev | 17:26 | |
| *** mjfork has quit IRC | 17:27 | |
| *** mjfork_ is now known as mjfork | 17:27 | |
| *** bauzas has joined #openstack-dev | 17:28 | |
| *** nati_ueno has quit IRC | 17:28 | |
| *** jayg has quit IRC | 17:28 | |
| *** nati_ueno has joined #openstack-dev | 17:29 | |
| *** gyee has joined #openstack-dev | 17:30 | |
| *** eglynn has quit IRC | 17:31 | |
| openstackstatus | NOTICE: Gerrit and Zuul going offline at 20:00 UTC for ~15mins for project renames | 17:34 |
| *** AlexF has joined #openstack-dev | 17:34 | |
| *** openstackstatus has quit IRC | 17:34 | |
| *** openstackstatus has joined #openstack-dev | 17:35 | |
| *** newell has joined #openstack-dev | 17:35 | |
| openstackstatus | NOTICE: Gerrit and Zuul going offline at 20:00 UTC for ~15mins for project renames | 17:35 |
| *** ChanServ changes topic to "Gerrit and Zuul going offline at 20:00 UTC for ~15mins for project renames" | 17:35 | |
| *** Longgeek has quit IRC | 17:36 | |
| *** gokrokve has joined #openstack-dev | 17:37 | |
| vishy | dolphm: if I want to get a project_name from a project_id and vice-versa using a service token | 17:38 |
| vishy | is there a good way to do that | 17:38 |
| vishy | secondarily is there a good way to do multiple at once? | 17:38 |
| *** chris_johnson is now known as wchrisj|away | 17:39 | |
| *** anniec has joined #openstack-dev | 17:39 | |
| *** florentflament has quit IRC | 17:40 | |
| *** akrivoka has joined #openstack-dev | 17:40 | |
| *** angdraug has joined #openstack-dev | 17:41 | |
| *** tanisdl has quit IRC | 17:41 | |
| *** gokrokve has quit IRC | 17:41 | |
| *** tqtran has joined #openstack-dev | 17:42 | |
| *** I159 has quit IRC | 17:42 | |
| *** tanisdl has joined #openstack-dev | 17:42 | |
| ayoung | vishy, right now the first is trivial; | 17:43 |
| ayoung | the second is not | 17:43 |
| ayoung | but multiple at once.... | 17:43 |
| ayoung | if you do a list projects, you get both, more data than you need, but maybe that is the right approach for your use | 17:44 |
| *** danpb has quit IRC | 17:44 | |
| *** ijw has quit IRC | 17:45 | |
| *** killer_prince has joined #openstack-dev | 17:46 | |
| *** david-lyle has joined #openstack-dev | 17:46 | |
| stevemar | bknudson, shoot, tests found an example where it wouldn't work unless we have all the data ready to replace | 17:46 |
| vishy | ayoung: the second meaning get an id from a name? | 17:47 |
| ayoung | vishy, yeah | 17:47 |
| *** SumitNaiksatam has joined #openstack-dev | 17:47 | |
| ayoung | there is an internal API to do that, but not exposed | 17:47 |
| ayoung | um, wait | 17:47 |
| *** nati_ueno has quit IRC | 17:48 | |
| ayoung | I think actually filter works for that. 1 sec | 17:48 |
| ayoung | get_project_by_name | 17:48 |
| dstanek | stevemar: still working on mapping issues? | 17:49 |
| ayoung | but that is v2 | 17:49 |
| stevemar | dstanek, you know it, just trying to make it awesome | 17:49 |
| ayoung | list_projects is what we want, takes a filter... | 17:49 |
| *** jcooley_ has joined #openstack-dev | 17:50 | |
| *** amerine has joined #openstack-dev | 17:51 | |
| dstanek | stevemar: nice | 17:51 |
| *** jp_at_hp has quit IRC | 17:52 | |
| *** Gordonz has joined #openstack-dev | 17:52 | |
| ayoung | vishy, so we have pretty complex logic for listing projects based on a user_id. | 17:52 |
| *** Gordonz has quit IRC | 17:52 | |
| *** jmckind has quit IRC | 17:52 | |
| *** max_lobur is now known as max_lobur_afk | 17:52 | |
| ayoung | We have an internal API get_project_by name, but that is only exposed via the V2 api | 17:53 |
| *** Gordonz has joined #openstack-dev | 17:53 | |
| *** markmcclain has quit IRC | 17:53 | |
| vishy | ayoung: so it is in v2 but not v3 | 17:53 |
| *** galstrom_zzz is now known as galstrom | 17:53 | |
| ayoung | vishy, still looking | 17:54 |
| ayoung | vishy, I thought we supported it with filters, too | 17:54 |
| ayoung | vishy, one thing that is wonky is we don';t have a concept of "this role means you get to know about projects inside this domain" | 17:54 |
| ayoung | we have list_projects_for_domain | 17:54 |
| ayoung | vishy, https://github.com/openstack/keystone/blob/master/keystone/assignment/routers.py | 17:55 |
| *** YorikSar has joined #openstack-dev | 17:56 | |
| ayoung | wow... | 17:56 |
| ayoung | we have some identity stuff in there | 17:56 |
| *** jcooley_ has quit IRC | 17:56 | |
| *** rodrigods has joined #openstack-dev | 17:57 | |
| *** rodrigods_ has joined #openstack-dev | 17:57 | |
| *** blues-man has quit IRC | 17:57 | |
| *** jprovazn_bbl has quit IRC | 17:59 | |
| *** marcoemorais has quit IRC | 17:59 | |
| ayoung | vishy, OK, starting from the API | 17:59 |
| ayoung | https://github.com/openstack/identity-api/blob/master/openstack-identity-api/v3/src/markdown/identity-api-v3.md#list-projects-get-projects | 17:59 |
| ayoung | It should be GET /v3/projects?name=<name> | 17:59 |
| *** marcoemorais has joined #openstack-dev | 18:00 | |
| ayoung | I assume that would required | 18:00 |
| ayoung | domain_id as well, to deconflict between two domains | 18:00 |
| *** bswartz has joined #openstack-dev | 18:00 | |
| stevemar | bknudson, the only change that i ended up making was returning early if the eval was false, and continuing if it was valid | 18:00 |
| ayoung | otherwise...I'm guessing it will match the domain from the user that is requesting it, but I have not tested it | 18:00 |
| *** marekd is now known as marekd|away | 18:00 | |
| ayoung | vishy, a lot of the logic is collected into the keystone/common/routers.py file | 18:02 |
| ayoung | that implementes the repeated CRUD operations | 18:02 |
| *** ijw has joined #openstack-dev | 18:03 | |
| ayoung | then the filters accepted would be passed down to the controller, in this case the V3 | 18:03 |
| *** jog0 is now known as flashgordon | 18:03 | |
| *** rossella_s has quit IRC | 18:04 | |
| ayoung | the majic is in here project_refs = self.filter_query(Project, query, hints) | 18:04 |
| *** BobBall is now known as BobBallAway | 18:04 | |
| *** cadenzajon has joined #openstack-dev | 18:05 | |
| *** andreaf has joined #openstack-dev | 18:06 | |
| ayoung | vishy, OK, that was way too "in the weeds" answer but the short of it is "yes, you can get a project object based on either id or name, with name being less efficient" I assume, though, that you are pursuing hierarchical with this... | 18:07 |
| *** rossella_s has joined #openstack-dev | 18:07 | |
| vishy | not exactly no | 18:08 |
| vishy | although it may apply | 18:08 |
| vishy | i was actually going to use it for dns | 18:08 |
| vishy | but i may use the sme thing in hierachical | 18:08 |
| *** kgriffs is now known as kgriffs_afk | 18:10 | |
| ayoung | so...you probably want to do a bulk list | 18:10 |
| *** lbragstad has quit IRC | 18:10 | |
| ayoung | and cache the results | 18:10 |
| *** lucasagomes has quit IRC | 18:11 | |
| ayoung | we send out notifications when a project is created, so you could register for that as well | 18:11 |
| ayoung | or just periodically refetch and rebuild | 18:11 |
| *** nacim has quit IRC | 18:11 | |
| ayoung | If we had a "parent" field in there (as opposed to just domain) you could, in theory, build the whole project name from parts by querying all projects for a domain | 18:12 |
| ayoung | and use that as the FQDN | 18:12 |
| *** martyntaylor has left #openstack-dev | 18:13 | |
| *** IanGovett has quit IRC | 18:13 | |
| *** nati_ueno has joined #openstack-dev | 18:14 | |
| *** AlexF has quit IRC | 18:14 | |
| *** sweston has quit IRC | 18:15 | |
| *** bauzas has quit IRC | 18:15 | |
| *** morganfainberg_Z is now known as morganfainberg | 18:15 | |
| *** dvarga has quit IRC | 18:16 | |
| *** jcooley_ has joined #openstack-dev | 18:16 | |
| *** lbragstad has joined #openstack-dev | 18:17 | |
| *** igor_ has quit IRC | 18:17 | |
| *** igor_ has joined #openstack-dev | 18:18 | |
| *** harlowja_away is now known as harlowja | 18:18 | |
| *** sarob has joined #openstack-dev | 18:18 | |
| *** sarob has quit IRC | 18:19 | |
| *** sarob has joined #openstack-dev | 18:19 | |
| *** jcooley_ has quit IRC | 18:20 | |
| *** otherwiseguy has quit IRC | 18:21 | |
| *** READ10 has quit IRC | 18:22 | |
| *** Mandell has joined #openstack-dev | 18:22 | |
| *** jasondotstar has joined #openstack-dev | 18:22 | |
| *** colinmcnamara has quit IRC | 18:23 | |
| *** AlexF has joined #openstack-dev | 18:23 | |
| *** igor_ has quit IRC | 18:24 | |
| *** e0ne has joined #openstack-dev | 18:24 | |
| *** galstrom is now known as galstrom_zzz | 18:26 | |
| *** markmcclain has joined #openstack-dev | 18:26 | |
| *** kushal has quit IRC | 18:27 | |
| *** wchrisj|away is now known as chris_johnson | 18:27 | |
| *** pablosan has joined #openstack-dev | 18:28 | |
| *** vuil has quit IRC | 18:28 | |
| *** capri has quit IRC | 18:28 | |
| *** capri has joined #openstack-dev | 18:31 | |
| *** mattymo has quit IRC | 18:33 | |
| *** AlexF has quit IRC | 18:34 | |
| *** mattymo has joined #openstack-dev | 18:34 | |
| *** jcooley_ has joined #openstack-dev | 18:40 | |
| *** RajeshMohan has quit IRC | 18:42 | |
| *** RajeshMohan has joined #openstack-dev | 18:43 | |
| *** mkollaro has quit IRC | 18:46 | |
| *** hemna has quit IRC | 18:46 | |
| *** hemna has joined #openstack-dev | 18:46 | |
| *** ijw has quit IRC | 18:47 | |
| *** zzelle has joined #openstack-dev | 18:47 | |
| *** hemna has quit IRC | 18:47 | |
| *** jcooley_ has quit IRC | 18:47 | |
| *** jcooley_ has joined #openstack-dev | 18:48 | |
| *** igor_ has joined #openstack-dev | 18:50 | |
| *** rossella_s has quit IRC | 18:51 | |
| *** dspano has joined #openstack-dev | 18:52 | |
| *** jcooley_ has quit IRC | 18:52 | |
| *** jcooley_ has joined #openstack-dev | 18:53 | |
| *** dvarga has joined #openstack-dev | 18:53 | |
| *** hemna has joined #openstack-dev | 18:55 | |
| *** igor_ has quit IRC | 18:55 | |
| *** anniec has quit IRC | 18:56 | |
| *** jamespage_ has joined #openstack-dev | 18:56 | |
| *** martyntaylor has joined #openstack-dev | 18:56 | |
| *** martyntaylor has left #openstack-dev | 18:56 | |
| *** jamespage_ has quit IRC | 18:58 | |
| openstackstatus | NOTICE: Zuul is now in queue-only mode preparing for project renames at 20:00 UTC | 18:59 |
| *** jayg|g0n` has joined #openstack-dev | 19:01 | |
| *** galstrom_zzz is now known as galstrom | 19:01 | |
| *** epim has joined #openstack-dev | 19:02 | |
| *** SumitNaiksatam has quit IRC | 19:02 | |
| *** jhesketh_ has quit IRC | 19:02 | |
| *** shakayumi has joined #openstack-dev | 19:02 | |
| *** melwitt has joined #openstack-dev | 19:03 | |
| *** anniec has joined #openstack-dev | 19:03 | |
| *** kenperkins_ is now known as kenperkins | 19:03 | |
| *** galstrom is now known as galstrom_zzz | 19:04 | |
| *** jcooley_ has quit IRC | 19:06 | |
| *** DinaBelova_ is now known as DinaBelova | 19:06 | |
| raildo | dolphm: you read the log from today's meeting about Hierarchical multitenancy? I and tellesnobrega were responsible for try making keystone have nested projects and send a bigger scope. I was thinking initially test a simple solution in which for a project I would create a column "parent_project" and populate it with the name(or id) of the parent project, the higher the level of this project and the project root it will be nul | 19:06 |
| dolphm | raildo: i did read it, yes | 19:06 |
| dolphm | raildo: only use id's internally, so it would be a parent_project_id | 19:07 |
| raildo | ok | 19:07 |
| *** thuc has quit IRC | 19:08 | |
| dolphm | raildo: are you looking to supersede domains using the project hierarchy? | 19:08 |
| *** thuc has joined #openstack-dev | 19:08 | |
| *** thuc has quit IRC | 19:08 | |
| raildo | no | 19:09 |
| *** thuc has joined #openstack-dev | 19:09 | |
| dolphm | raildo: you don't think that domains should become 1st level projects, after a root project? | 19:10 |
| *** thuc has quit IRC | 19:10 | |
| *** thuc_ has joined #openstack-dev | 19:10 | |
| raildo | I want to follow the idea that there would be a domain, and root project attached to it, and then create the project hierarchy. | 19:10 |
| *** thuc_ has quit IRC | 19:11 | |
| *** thuc has joined #openstack-dev | 19:11 | |
| raildo | dolphm: IMO, domain is very important for the architecture of the infrastructure in openstack. | 19:12 |
| *** jcooley_ has joined #openstack-dev | 19:13 | |
| raildo | Only it was not clear to me, the issue of increasing the scope of the token but will primarily focus on the implementation of nested project. | 19:15 |
| *** johnthetubaguy has quit IRC | 19:15 | |
| *** arnaud has joined #openstack-dev | 19:16 | |
| *** arnaud__ has joined #openstack-dev | 19:16 | |
| *** annashen has left #openstack-dev | 19:16 | |
| *** amcrn has joined #openstack-dev | 19:18 | |
| *** Longgeek has joined #openstack-dev | 19:18 | |
| *** kgriffs_afk is now known as kgriffs | 19:18 | |
| dolphm | raildo: so in your perspective, each domain has a 1:1 relationship with a "root" project? | 19:22 |
| morganfainberg | raildo, dolphm, i think that is a ... bad structure | 19:22 |
| dolphm | morganfainberg: ++ | 19:22 |
| morganfainberg | dare i say, pointless | 19:22 |
| dolphm | post-migration: http://i.imgur.com/vBsrdi2.png | 19:22 |
| morganfainberg | i'd advocate domains existing as an abstraction if we want to keep user-namespacing in them | 19:22 |
| *** Longgeek has quit IRC | 19:23 | |
| dolphm | migrating all domains to be project will null parents, and migrating all projects to be exposed as children of the new top-level projects | 19:23 |
| dolphm | morganfainberg: +++ | 19:23 |
| morganfainberg | dolphm, in fact, that resolves my complaints about domain issues. | 19:23 |
| morganfainberg | dolphm, if projects no-longer are "in domains" but are just (effectively) in other projects | 19:24 |
| dolphm | yep | 19:24 |
| *** galstrom_zzz is now known as galstrom | 19:24 | |
| morganfainberg | conceptually at least it makes it easier to talk about... and all that stuff | 19:24 |
| *** Mandell has quit IRC | 19:24 | |
| *** sweston has joined #openstack-dev | 19:25 | |
| raildo | I had thought about a 1:1 initially had not thought this way, | 19:26 |
| *** sweston has quit IRC | 19:26 | |
| *** gordc has quit IRC | 19:26 | |
| *** sweston has joined #openstack-dev | 19:26 | |
| raildo | I think the structure will be so much better | 19:26 |
| *** rcleere has joined #openstack-dev | 19:27 | |
| morganfainberg | raildo, pointless data structures for the sake of pointless data structures is ugly and leads to legacy code ;) | 19:27 |
| raildo | +1 | 19:27 |
| dolphm | morganfainberg: how do we do this migration without breaking every other project? add an option to auth_token that mutates the project scope with X-PROJECT-ID: project_id.split('.').pop() ? | 19:29 |
| morganfainberg | dolphm, hmmmmm. | 19:29 |
| *** SumitNaiksatam has joined #openstack-dev | 19:29 | |
| tellesnobrega | tellesnobrega: i might have come too late to this conversation, why should we have a root_project with domains and then projects again attached to the domain? | 19:29 |
| dolphm | [auth_token] do_not_understand_hierarchical_multitenancy = true | 19:29 |
| morganfainberg | dolphm, LOL. | 19:29 |
| dolphm | seriously! | 19:29 |
| morganfainberg | dolphm, can we just add a new variable that they can consume when they want? | 19:30 |
| *** nati_uen_ has joined #openstack-dev | 19:30 | |
| morganfainberg | dolphm, X-PROJECT-HIERARCHY | 19:30 |
| *** jamieh has quit IRC | 19:30 | |
| morganfainberg | or something | 19:30 |
| dolphm | morganfainberg: HMMMMMMMMMMMMMMMMMMMMMMMMMMM | 19:30 |
| dolphm | morganfainberg: -ID is forever a lie then | 19:30 |
| *** nati_uen_ has quit IRC | 19:30 | |
| morganfainberg | dolphm, it is likely the easiest migration path. offer both, and deprecate out the old one in X-cycles | 19:30 |
| morganfainberg | dolphm, it might be a lie, but it is no different than someone trying to consume V2 API vs V3. | 19:31 |
| *** nati_uen_ has joined #openstack-dev | 19:31 | |
| morganfainberg | V2 is ... going to give you a wildly different answer in some cases | 19:31 |
| morganfainberg | and if they don't care about the hierarchy initially, does it really matter? they wont be setup for the cascading roles etc anyway | 19:32 |
| *** nati_ueno has quit IRC | 19:32 | |
| tellesnobrega | morganfainberg, dolphm the point of having the root_project on top of domains is to offer the super admin role or is there another use for it? | 19:33 |
| *** thuc has quit IRC | 19:33 | |
| dolphm | morganfainberg: yeah, we'd have to shield v2 from the impact of this quite a lot | 19:33 |
| morganfainberg | tellesnobrega, there is no difference between a root project or a domain | 19:34 |
| dolphm | tellesnobrega: that's the primary use case in my mind | 19:34 |
| morganfainberg | tellesnobrega, in that context really | 19:34 |
| *** thuc has joined #openstack-dev | 19:34 | |
| *** lyxus has quit IRC | 19:34 | |
| morganfainberg | short of having a domain that can contain root projects... | 19:34 |
| morganfainberg | which, seems to defeat the point. | 19:34 |
| dolphm | tellesnobrega: i think vishy suggested that that project have an id of 'openstack' | 19:34 |
| tellesnobrega | dolphm: ++ | 19:34 |
| dolphm | morganfainberg: which would actually be better for our role assignment api ^ | 19:34 |
| morganfainberg | dolphm, i actually like that | 19:34 |
| *** anniec has quit IRC | 19:35 | |
| morganfainberg | dolphm, we make an implicit (and by implicit i mean code construct) project called Openstack. | 19:35 |
| *** hartsocks is now known as hartbot | 19:35 | |
| dolphm | morganfainberg: then every project ID becomes something like openstack.default.e5f9y8jdygdz95wpu3e5wuj3t | 19:35 |
| *** lyxus has joined #openstack-dev | 19:35 | |
| morganfainberg | dolphm, i don't want to have to seed the project into the DB if we have a "root" project like that | 19:35 |
| dolphm | morganfainberg: if it exists in the db, then objects have to explicitly point to it, and then we'll have users that want to create a second tree | 19:36 |
| tellesnobrega | morganfainberg: ++ | 19:36 |
| morganfainberg | dolphm, that is why i think it should be a code construct -- there isn't duplicate trees | 19:36 |
| tellesnobrega | the id should be like domain1.kl2jhkl1jhkl25jkl12j5 | 19:36 |
| dolphm | morganfainberg: yay | 19:36 |
| morganfainberg | dolphm, we can then move domains into being the management for IDPs if we want to keep that term (since we already use it in the REST api) | 19:37 |
| dolphm | morganfainberg: aaand you lost me | 19:37 |
| raildo | dolphm: tellesnobrega and I'll start implementing it =] | 19:37 |
| morganfainberg | dolphm, domains no longer are an assignment construct | 19:38 |
| dolphm | morganfainberg: still lost | 19:38 |
| morganfainberg | dolphm, they are a user-namespace management construct | 19:38 |
| dolphm | morganfainberg: "domain" == "project with null parent" ? | 19:38 |
| *** adnan_ has joined #openstack-dev | 19:38 | |
| morganfainberg | dolphm, nah, projects would cease to care about domains | 19:38 |
| *** thuc has quit IRC | 19:38 | |
| morganfainberg | dolphm, root project = openstack | 19:38 |
| *** mrodden has quit IRC | 19:38 | |
| morganfainberg | there is no domain container really. | 19:38 |
| morganfainberg | don't even call it a domain :P | 19:39 |
| morganfainberg | no reason to | 19:39 |
| dolphm | morganfainberg: right, but in terms of GET /v3/domains, why not return SELECT * FROM projects WHERE parent_project_id IS NULL; ? | 19:39 |
| morganfainberg | dolphm, ehhhhh | 19:39 |
| dolphm | i know | 19:39 |
| *** vkmc has joined #openstack-dev | 19:39 | |
| dolphm | morganfainberg: you just want to kill GET /v3/domains immediately? | 19:39 |
| morganfainberg | dolphm, i'd really like to move us away from referring to it as domains sure for transitional | 19:39 |
| dolphm | i'm just talking juno | 19:40 |
| *** adnan_ has left #openstack-dev | 19:40 | |
| morganfainberg | dolphm, well, it is a question of if domains will stick around for user stuff in V3 | 19:40 |
| morganfainberg | dolphm, i am guessing it has to for Idenitty purposes | 19:40 |
| morganfainberg | dolphm, short of moving to V4 | 19:41 |
| morganfainberg | since "domain" still constructs the wrapper for authentication for users and the like... | 19:41 |
| *** AnilV4 has quit IRC | 19:41 | |
| *** LLKCKfan has joined #openstack-dev | 19:41 | |
| LLKCKfan | Is there any natural ways to relieve pain without using herbs or weed? No drugs | 19:41 |
| morganfainberg | dolphm, so, do we just go w/ domains being parentless projects | 19:42 |
| morganfainberg | oh, joy a spammer.... | 19:42 |
| *** thuc has joined #openstack-dev | 19:42 | |
| dolphm | ttx: jeblair: ^ | 19:42 |
| morganfainberg | didn't this guy get banned in the pas? | 19:42 |
| morganfainberg | past* | 19:42 |
| raildo | hahahaha | 19:42 |
| dolphm | i don't recognize the name | 19:42 |
| morganfainberg | mordred, ^ | 19:42 |
| *** mrodden has joined #openstack-dev | 19:43 | |
| tellesnobrega | shouldn't domains have parent_project_id = openstack? | 19:44 |
| tellesnobrega | or did i lose something in the way? | 19:44 |
| dolphm | tellesnobrega: i'd like to do that at the controller layer | 19:44 |
| *** julienvey has quit IRC | 19:44 | |
| dolphm | or even managers | 19:45 |
| morganfainberg | tellesnobrega, the "openstack" parent_project needs to be a code construct not an element in the db | 19:45 |
| *** alexpilotti has quit IRC | 19:45 | |
| *** SumitNaiksatam has quit IRC | 19:45 | |
| morganfainberg | dolphm, ++ managers might be the right place | 19:45 |
| tellesnobrega | i see | 19:45 |
| *** Mandell has joined #openstack-dev | 19:45 | |
| marun | anybody able to answer questions about docker in openstack? | 19:48 |
| *** igor_ has joined #openstack-dev | 19:49 | |
| dolphm | morganfainberg: so maybe since we're dumping domains, we should go back to v2 -- that's how that works, right? | 19:49 |
| morganfainberg | dolphm, lol | 19:49 |
| *** sarob has quit IRC | 19:50 | |
| dolphm | "Due to support for hierarchical multitenancy, v3 is now deprecated in favor of v2. Hugs and kisses, -keystone" | 19:50 |
| morganfainberg | dolphm, i vote we go back to pre-V2 days | 19:50 |
| *** ChanServ sets mode: +o jeblair | 19:50 | |
| dolphm | morganfainberg: reboot v1? | 19:50 |
| ayoung | morganfainberg, food for thought: a project "belongs" to a domain in a way that it will not "belong" to a parent project.... | 19:50 |
| morganfainberg | dolphm, ++ | 19:50 |
| *** jeblair sets mode: +b *!~LLKCKfan@* | 19:50 | |
| ayoung | IE | 19:50 |
| *** vartom1111111117 has joined #openstack-dev | 19:50 | |
| marun | samalba: ping | 19:50 |
| dolphm | jeblair: thanks! | 19:50 |
| morganfainberg | ayoung, sorry i don't use IE | 19:51 |
| ayoung | if I want to move a project from one parent to another, good to go | 19:51 |
| *** LLKCKfan was kicked by jeblair (LLKCKfan) | 19:51 | |
| morganfainberg | oooh i.e. | 19:51 |
| dolphm | ayoung: that changes the project's exposed ID, and breaks every other project | 19:51 |
| ayoung | e.i.e.i.o | 19:51 |
| dolphm | ayoung: parent_project_id has to immutable | 19:51 |
| morganfainberg | ayoung, i think part of the discussion was projects do not get reparented | 19:51 |
| morganfainberg | dolphm, but people will ask for that >.< | 19:52 |
| * dolphm what if projects only existed as assignment targets | 19:52 | |
| ayoung | but...that is a an easy answer for Keystone, but tough for people that use projects to contain *things* | 19:52 |
| dolphm | ayoung: things can move | 19:52 |
| dolphm | ayoung: that's up to the other services | 19:53 |
| *** clu_ has joined #openstack-dev | 19:53 | |
| ayoung | I think that people will want to move projects around in the hierarchy | 19:53 |
| ayoung | so if a VM has a project_id for "owner" or whatever | 19:53 |
| dolphm | ayoung: then move the VM to a new project | 19:53 |
| dolphm | reparent individual resources, rather than breaking the entire cloud | 19:53 |
| ayoung | dolphm, that is like saying "hey Nova, do all the tough work" | 19:53 |
| ayoung | and then you need to keep it in sync across projects | 19:53 |
| ayoung | nah, no me likey | 19:53 |
| dolphm | ayoung: keep what in sync where? | 19:54 |
| ayoung | we need to be able to reparent projects without breaking things | 19:54 |
| dolphm | ayoung: that's literally impossible | 19:54 |
| ayoung | dolphm, network for a VM between nova nad Neutron? | 19:54 |
| *** dvarga is now known as dvarga|away | 19:54 | |
| *** dvarga|away is now known as dvarga | 19:54 | |
| morganfainberg | ayoung, this is a cloud, if you need a resource in another project, spin one up :P | 19:55 |
| *** anniec has joined #openstack-dev | 19:55 | |
| ayoung | dolphm, morganfainberg do you understand how SELinux lables inodes? | 19:55 |
| ayoung | I would say this: | 19:55 |
| morganfainberg | ayoung, ok ok now that i got that out of my system | 19:55 |
| ayoung | domain_id stays immutable for a project | 19:55 |
| ayoung | but projects are like dentries | 19:55 |
| ayoung | and I should be able to move a dentry from one parent to another and carry along the contents | 19:56 |
| ayoung | but not from one domain to another | 19:56 |
| dolphm | morganfainberg: lol | 19:56 |
| ayoung | then, something like a VM can have a project_id, and , if that project moves throughout the tree, fine | 19:56 |
| dolphm | ayoung: domains don't exist anymore, and are just an arbitrarily constrained perspective on the project hierarchy | 19:57 |
| ayoung | its project id is immutable, but the absolute project name may change...to ditinguish from a relative project name... parent/child vs child | 19:57 |
| *** ijw has joined #openstack-dev | 19:57 | |
| morganfainberg | ayoung, can we table that and come back around to solving it as we have demand for it? | 19:58 |
| ayoung | morganfainberg, no no no no no | 19:58 |
| ayoung | we punt on this and they get it wrong, everyone suffers | 19:58 |
| morganfainberg | ayoung, i think that is adding a lot of scope creep, and making this a hard target for juno | 19:58 |
| ayoung | no | 19:58 |
| ayoung | this is getting the abstraction right | 19:58 |
| ayoung | I'm ok with getting rid of domains. | 19:58 |
| *** alexpilotti has joined #openstack-dev | 19:58 | |
| *** dprince has quit IRC | 19:59 | |
| * annegentle waves at dims | 19:59 | |
| morganfainberg | moving a project is in-dependant of the hierarchical structure | 19:59 |
| ayoung | just wanted to make sure it is a deliberate decision and we understand the implications | 19:59 |
| ayoung | morganfainberg, project ID id immutable. parent_id does not have to be | 19:59 |
| morganfainberg | ayoung, it can be to start, and then we can add that capability in | 19:59 |
| ayoung | lets not force it to be until we have some stake in the ground reason | 20:00 |
| ayoung | I would say "domain is a project with no parent" | 20:00 |
| *** julienvey has joined #openstack-dev | 20:00 | |
| ayoung | that way we provide language for people to port from the existing docs | 20:00 |
| ayoung | or | 20:00 |
| ayoung | domains are projects that live directly under the root node | 20:00 |
| *** MaxV has joined #openstack-dev | 20:01 | |
| morganfainberg | ayoung, i'd rather have a solid understanding of the hierarchy and present that (and associated roles, etc) to the services and then figure out how to "move" things rather than figure out how to move things then define the hierarchy | 20:01 |
| morganfainberg | ayoung, if that makes sense. | 20:01 |
| *** MaxV has quit IRC | 20:01 | |
| ayoung | morganfainberg, so I am not the one that originated this idea...it has come up before... | 20:01 |
| ayoung | trying to drag the conversation out of long term memory... | 20:01 |
| morganfainberg | you can say avoid staking in the ground "partents are immutable" but you also don't provide a way to move things. | 20:01 |
| *** ijw has quit IRC | 20:02 | |
| *** thuc has quit IRC | 20:02 | |
| ayoung | PATCH /v3/projects/<id> {parente=new id} | 20:02 |
| morganfainberg | but.. you're going to have other issues when it comes to keeping all that in sync w/ nova etc as well | 20:02 |
| ayoung | we'd have to explicitly forbit that | 20:02 |
| *** thuc has joined #openstack-dev | 20:03 | |
| morganfainberg | ayoung, we would want to forbid that in a perfect world anyway | 20:03 |
| morganfainberg | ayoung, i'd argue that should be a separate API in either case | 20:03 |
| ayoung | its an update notification. | 20:03 |
| morganfainberg | reparenting has a much much larger implication than other updates | 20:03 |
| ayoung | why complicate things | 20:03 |
| *** gokrokve has joined #openstack-dev | 20:03 | |
| morganfainberg | ayoung, because it isn't just an update. | 20:03 |
| ayoung | no, disable has much larger implications | 20:03 |
| *** gokrokve has quit IRC | 20:03 | |
| *** gokrokve has joined #openstack-dev | 20:03 | |
| morganfainberg | ayoung, reparent has more incommon with disable | 20:04 |
| morganfainberg | ayoung, you could be forcing a masiive shift in permissions / access | 20:04 |
| *** thuc has quit IRC | 20:04 | |
| morganfainberg | ayoung, than a project getting an update. | 20:04 |
| *** thuc has joined #openstack-dev | 20:04 | |
| ayoung | morganfainberg, the person making the change would need a role on both the source and target parent projects. | 20:04 |
| dolphm | morganfainberg: ayoung: raildo: tellesnobrega: vishy: starting drafting the long term impact on keystone https://blueprints.launchpad.net/keystone/+spec/hierarchical-multitenancy | 20:04 |
| ayoung | create subproject | 20:05 |
| ayoung | and delete subproject | 20:05 |
| raildo | dolphm: i just sign up on it, thanks | 20:06 |
| tellesnobrega | tellesnobrega: me too | 20:06 |
| dolphm | tellesnobrega: raildo: thanks, i'm sure we'll all be pitching in | 20:06 |
| *** henrique has joined #openstack-dev | 20:07 | |
| *** capri has quit IRC | 20:07 | |
| *** tjones has quit IRC | 20:07 | |
| tellesnobrega | dolphm: we want to have something done by friday so we can show vishy at the meeting, of course we won't be able to have all that was discussed done, but we would like to have the hierarchy working | 20:08 |
| *** akrivoka has quit IRC | 20:08 | |
| *** capri has joined #openstack-dev | 20:08 | |
| *** ewindisch has joined #openstack-dev | 20:09 | |
| *** jruzicka has quit IRC | 20:09 | |
| raildo | +1 | 20:09 |
| *** ewindisch has quit IRC | 20:10 | |
| openstackstatus | NOTICE: Gerrit and Zuul are offline for project renames. ETA 20:30 UTC. | 20:10 |
| *** ChanServ changes topic to "Gerrit and Zuul are offline for project renames. ETA 20:30 UTC." | 20:10 | |
| tellesnobrega | and write something to submit for the summit and present it there. | 20:10 |
| *** ewindisch has joined #openstack-dev | 20:10 | |
| vishy | ayoung: cool | 20:12 |
| vishy | dolphm: ^^ | 20:12 |
| tellesnobrega | vishy: raildo and I are going to implement this hierarchical projects, or whatever we can do of it, by friday so we can show at the multitenancy meeting and bring it to the summit, did you get a chance to read this whole conversation? | 20:13 |
| *** networkstatic has joined #openstack-dev | 20:16 | |
| *** edmund has quit IRC | 20:17 | |
| *** anniec has quit IRC | 20:18 | |
| *** otherwiseguy has joined #openstack-dev | 20:18 | |
| *** DinaBelova is now known as DinaBelova_ | 20:18 | |
| *** comstud is now known as bearhands | 20:18 | |
| *** rfolco has quit IRC | 20:18 | |
| ayoung | dolphm, geting the client review done early is going to be crucial for the token compression. I appreciate it if you rip it apart soon-rather-than-later. https://review.openstack.org/#/c/71181/ | 20:19 |
| *** Longgeek has joined #openstack-dev | 20:19 | |
| *** nati_ueno has joined #openstack-dev | 20:19 | |
| ayoung | dstanek, bknudson ^^ | 20:19 |
| dolphm | ayoung: the referenced bug is not tracked against the client | 20:19 |
| dolphm | ayoung: and gerrit is down :) | 20:20 |
| bknudson | 503 Service Temporarily Unavailable | 20:20 |
| *** DinaBelova_ is now known as DinaBelova | 20:20 | |
| dolphm | from fridge import beer; | 20:21 |
| *** thuc has quit IRC | 20:21 | |
| morganfainberg | dolphm, from kegerator import tap \n pint = tap.pour_beer pint.drink() | 20:21 |
| *** thuc has joined #openstack-dev | 20:21 | |
| dolphm | morganfainberg: SyntaxError ^ | 20:22 |
| morganfainberg | dolphm, LOL | 20:22 |
| *** nati_ueno has quit IRC | 20:22 | |
| *** taps has quit IRC | 20:22 | |
| *** nati_uen_ has quit IRC | 20:22 | |
| *** tjones has joined #openstack-dev | 20:23 | |
| *** nati_ueno has joined #openstack-dev | 20:23 | |
| *** dbalog has left #openstack-dev | 20:23 | |
| ayoung | dolphm, FTFY | 20:23 |
| ayoung | https://bugs.launchpad.net/python-keystoneclient/+bug/1255321 | 20:23 |
| *** Longgeek has quit IRC | 20:24 | |
| uvirtbot | Launchpad bug 1255321 in keystone "v3 token requests result in 500 error when run in apache" [Medium,Confirmed] | 20:24 |
| *** taps has joined #openstack-dev | 20:24 | |
| dolphm | ayoung: danke | 20:24 |
| ayoung | Bitte | 20:24 |
| *** eglynn has joined #openstack-dev | 20:25 | |
| dolphm | morganfainberg: http://pasteraw.com/64uboev8qhk7cfggv4yejjurrn18vdq | 20:25 |
| * ayoung going to gym to perform 1/lunch | 20:25 | |
| *** MaxV has joined #openstack-dev | 20:25 | |
| morganfainberg | dolphm, http://pasteraw.com/4hmo6g7u0ivdqhakqn80e3qj9wif59d | 20:25 |
| morganfainberg | better? | 20:26 |
| morganfainberg | :P | 20:26 |
| *** ChanServ changes topic to "OpenStack development || Support is in #openstack" | 20:26 | |
| *** thuc has quit IRC | 20:26 | |
| ayoung | morganfainberg, the pur_craft_beer function makes no sense, you need to select the craft beer tap from the taps collection | 20:26 |
| morganfainberg | ayoung, nah, i have the simplified keggerator library | 20:27 |
| *** vkmc has quit IRC | 20:27 | |
| ayoung | otherwise, you get whatever keggerator has | 20:27 |
| *** tjones has quit IRC | 20:27 | |
| ayoung | so just "pour" | 20:27 |
| ayoung | -1 | 20:27 |
| morganfainberg | you're thinking of the bar or tap room mechanism | 20:27 |
| *** edmund has joined #openstack-dev | 20:27 | |
| morganfainberg | since the non-simplified library is not in globale requirements, we can't use it yet | 20:27 |
| morganfainberg | and the collection is unavailable | 20:27 |
| ayoung | import bar | 20:27 |
| ayoung | galss.contensts = bar.taps['Guinness'].pour | 20:28 |
| dolphm | morganfainberg: success http://pasteraw.com/hh9vbloy84o11xo1tb9vcvsosdpyjos | 20:28 |
| ayoung | glass.contensts = bar.taps['Guinness'].pour | 20:28 |
| ayoung | ugh...fuergetit me go gym | 20:28 |
| dolphm | morganfainberg: although this apparently works too http://pasteraw.com/bw5bxwu2xrq4pjqr1avykeoihq5q0m4 | 20:29 |
| dolphm | never tried piping straight to python .. | 20:29 |
| bknudson | curl | python -mjson.tool | 20:29 |
| dolphm | bknudson: but the module is just being executed there, not the curl'd data | 20:30 |
| ayoung | sometimes I wish they had chosen python instead of Javascript for the browser language. And by sometimes, I mean daytime and nightimes. | 20:30 |
| bknudson | I think you just reinvented java applets | 20:30 |
| morganfainberg | dolphm, https://twitter.com/MdrnStm/status/431887647242342401 | 20:30 |
| dolphm | bknudson: lol | 20:30 |
| dolphm | ayoung: i assume .pour was supposed to be a callable | 20:31 |
| morganfainberg | dolphm, nah it's a true/false :P | 20:32 |
| dolphm | ayoung: unless you meant .poor which is a @property of self | 20:32 |
| morganfainberg | lol | 20:32 |
| morganfainberg | if self.poor: bar.exit_sober() ? | 20:32 |
| *** sweston has quit IRC | 20:32 | |
| raildo | about this conversation http://ryanesaki.com/wp-content/uploads/2014/01/Anchorman_well_that_escalated_quickly_966.jpg hahahahaha | 20:32 |
| *** READ10 has joined #openstack-dev | 20:33 | |
| dolphm | morganfainberg: should definitely be a raise | 20:33 |
| morganfainberg | dolphm, assert self.poor is False | 20:33 |
| morganfainberg | or is that inverted | 20:33 |
| dolphm | morganfainberg: raise PaymentProcessorError() ? | 20:33 |
| morganfainberg | LOL | 20:33 |
| dolphm | assert not self.poor # reads a bit better | 20:34 |
| morganfainberg | except PaymentProcessError: self.dine_and_dash() | 20:34 |
| dolphm | lol | 20:34 |
| *** igor___ has joined #openstack-dev | 20:34 | |
| *** csd has quit IRC | 20:35 | |
| dolphm | morganfainberg: dining belongs in the try block http://pasteraw.com/tf1arys1o4xe6b2p5uiwhoj3d6d0j0y | 20:35 |
| morganfainberg | dolphm, hehe | 20:35 |
| dolphm | oh, gerrit is back | 20:36 |
| *** igor_ has quit IRC | 20:36 | |
| morganfainberg | yeah has been for a few minutes | 20:36 |
| *** anniec has joined #openstack-dev | 20:36 | |
| dstanek | ayoung: i don't think you addressed any of my comments yet | 20:36 |
| *** Longgeek has joined #openstack-dev | 20:37 | |
| bknudson | ever seen this? https://pypi.python.org/pypi/Spawning/ | 20:37 |
| bknudson | was just looking at eventlet docs | 20:37 |
| dolphm | dstanek: yep | 20:37 |
| *** nati_ueno has quit IRC | 20:38 | |
| devoid | anyone here have experience getting nova to use sheepdog root disks? | 20:38 |
| dolphm | why are we supporting two PKI token formats? | 20:38 |
| dolphm | why not read both, but only produce the compressed format? | 20:38 |
| *** tjones has joined #openstack-dev | 20:38 | |
| dolphm | i know i've asked this twice, but i don't recall ayoung answering ^ | 20:39 |
| *** markmcclain has quit IRC | 20:39 | |
| *** Longgeek has quit IRC | 20:41 | |
| *** nati_ueno has joined #openstack-dev | 20:41 | |
| dstanek | dolphm: i thought it was to support old clients | 20:41 |
| dolphm | dstanek: old keystoneclients? | 20:41 |
| *** jasondotstar has quit IRC | 20:42 | |
| *** nati_ueno has quit IRC | 20:42 | |
| *** anniec has quit IRC | 20:42 | |
| bknudson | old auth_token middlewares? | 20:42 |
| *** nati_ueno has joined #openstack-dev | 20:42 | |
| *** NikitaKonovalov_ is now known as NikitaKonovalov | 20:43 | |
| *** tjones has quit IRC | 20:43 | |
| dolphm | bknudson: so we want to support the combination of a new client in keystone, and old clients in front of other services? | 20:43 |
| dolphm | doesn't seem like it's worth all this duplicated crap | 20:44 |
| bknudson | dolphm: that would be a weird combo... not sure why anyone would be stuck with that... maybe a distributed upgrade? | 20:44 |
| dolphm | bknudson: even then -- upgrade auth_token middlewares to the latest client first -- done. | 20:44 |
| dolphm | i *really* don't want to ship 3 token formats and have to explain that one of them has no use case | 20:45 |
| dstanek | dolphm: ++ | 20:45 |
| sdague | david-lyle: I wanted to circle around with you again about the failure with horizon if we set -o errexit in devstack | 20:46 |
| *** rfolco has joined #openstack-dev | 20:46 | |
| *** cadenzajon has quit IRC | 20:47 | |
| *** thomasem has quit IRC | 20:48 | |
| *** tjones has joined #openstack-dev | 20:49 | |
| *** pingveno has quit IRC | 20:49 | |
| dolphm | -2'd for now | 20:50 |
| *** denis_makogon_ has joined #openstack-dev | 20:51 | |
| *** sarob has joined #openstack-dev | 20:52 | |
| *** ijw has joined #openstack-dev | 20:56 | |
| *** CaptTofu has quit IRC | 20:57 | |
| *** tjones has quit IRC | 20:57 | |
| *** sarob has quit IRC | 20:57 | |
| *** pablosan has quit IRC | 20:59 | |
| *** lcheng_ has joined #openstack-dev | 21:00 | |
| *** rcleere has quit IRC | 21:01 | |
| *** mgagne has quit IRC | 21:01 | |
| *** anniec has joined #openstack-dev | 21:03 | |
| *** sarob has joined #openstack-dev | 21:06 | |
| *** safchain has joined #openstack-dev | 21:06 | |
| *** prad_ has quit IRC | 21:06 | |
| *** safchain has quit IRC | 21:06 | |
| *** sarob has quit IRC | 21:10 | |
| *** godara has quit IRC | 21:10 | |
| *** IanGovett has joined #openstack-dev | 21:10 | |
| *** dbalog has joined #openstack-dev | 21:11 | |
| *** jgrimm has quit IRC | 21:11 | |
| *** jgrimm has joined #openstack-dev | 21:12 | |
| *** cadenzajon has joined #openstack-dev | 21:13 | |
| *** doug_shelley66 has quit IRC | 21:13 | |
| *** rfolco has quit IRC | 21:13 | |
| *** doug_shelley66 has joined #openstack-dev | 21:14 | |
| stevemar | lbragstad, good detective skills on https://review.openstack.org/#/c/66642/ | 21:16 |
| lbragstad | :) hopefully it works | 21:16 |
| *** yolanda has quit IRC | 21:18 | |
| *** DinaBelova is now known as DinaBelova_ | 21:19 | |
| *** tjones has joined #openstack-dev | 21:20 | |
| *** Longgeek has joined #openstack-dev | 21:20 | |
| *** byeager has quit IRC | 21:21 | |
| *** thuc has joined #openstack-dev | 21:21 | |
| *** Longgeek has quit IRC | 21:21 | |
| *** jmeridth has left #openstack-dev | 21:21 | |
| *** byeager has joined #openstack-dev | 21:22 | |
| yjiang5 | baoli: hi | 21:22 |
| *** lcheng_ has quit IRC | 21:22 | |
| *** Longgeek has joined #openstack-dev | 21:23 | |
| *** eglynn has quit IRC | 21:23 | |
| *** Longgeek has quit IRC | 21:23 | |
| *** tjones has quit IRC | 21:23 | |
| *** Longgeek has joined #openstack-dev | 21:24 | |
| *** Longgeek has quit IRC | 21:25 | |
| *** byeager has quit IRC | 21:25 | |
| *** Longgeek has joined #openstack-dev | 21:25 | |
| *** Longgeek has quit IRC | 21:26 | |
| *** Longgeek has joined #openstack-dev | 21:26 | |
| *** jayg|g0n` has quit IRC | 21:27 | |
| *** Longgeek has quit IRC | 21:27 | |
| *** Longgeek has joined #openstack-dev | 21:28 | |
| *** Longgeek has quit IRC | 21:29 | |
| *** Longgeek has joined #openstack-dev | 21:29 | |
| *** Longgeek has quit IRC | 21:30 | |
| *** networkstatic has quit IRC | 21:31 | |
| *** networkstatic has joined #openstack-dev | 21:32 | |
| *** Longgeek has joined #openstack-dev | 21:32 | |
| *** anniec has quit IRC | 21:32 | |
| *** Longgeek has quit IRC | 21:32 | |
| *** Longgeek has joined #openstack-dev | 21:33 | |
| *** sweston has joined #openstack-dev | 21:33 | |
| *** Longgeek has quit IRC | 21:33 | |
| *** doug-fish has quit IRC | 21:36 | |
| *** Longgeek has joined #openstack-dev | 21:36 | |
| *** Longgeek has quit IRC | 21:36 | |
| *** Longgeek has joined #openstack-dev | 21:38 | |
| *** doug-fish has joined #openstack-dev | 21:38 | |
| *** xqueralt has quit IRC | 21:38 | |
| *** Longgeek has quit IRC | 21:39 | |
| *** Longgeek has joined #openstack-dev | 21:39 | |
| stevemar | bknudson, i'm reverting the name changes from RULE_blah to MAPPING_blah. reasoning is, I want them the fixtures to be used for the mappingCRUD tests and the rule processing tests, either way I'm going to have to either grab the rules out of the mapping, or pad them with a {mapping:{rules: []}} block | 21:40 |
| *** Longgeek has quit IRC | 21:40 | |
| *** tmclaugh[work] has quit IRC | 21:40 | |
| bknudson | stevemar: that makes sense... if they're mappings call them mappings and if they're rules then call them rules. | 21:40 |
| stevemar | bknudson, deal | 21:40 |
| bknudson | not sure why we have to use the same fixtures for rule tests and mapping tests. | 21:40 |
| *** Longgeek has joined #openstack-dev | 21:40 | |
| bknudson | or why the fixtures have to be in a separate file... just put them in the test class. | 21:41 |
| *** anniec has joined #openstack-dev | 21:41 | |
| bknudson | I shouldn't have to go digging around to separate files to figure out what the test is. | 21:41 |
| *** jasondotstar has joined #openstack-dev | 21:41 | |
| *** Longgeek has quit IRC | 21:41 | |
| *** gordc has joined #openstack-dev | 21:41 | |
| *** Longgeek has joined #openstack-dev | 21:42 | |
| *** Longgeek has quit IRC | 21:42 | |
| *** Longgeek has joined #openstack-dev | 21:43 | |
| bknudson | stevemar: seems like there should be a test that does rp = mapping_utils.RuleProcessor({}) -- shouldn't have to go to a separate file to find that rules is empty. | 21:43 |
| *** glenng has quit IRC | 21:44 | |
| *** Longgeek has quit IRC | 21:44 | |
| bknudson | RULES_SMALL doesn't tell me anything about what's being tested in the rules processor | 21:44 |
| *** Longgeek has joined #openstack-dev | 21:44 | |
| *** luisbg has left #openstack-dev | 21:45 | |
| *** Longgeek has quit IRC | 21:45 | |
| *** mrodden has quit IRC | 21:46 | |
| stevemar | bknudson, I can do the first change in another commit, moving the fixture around, but it'll only be tested in MappingCRUD | 21:46 |
| *** tjones has joined #openstack-dev | 21:47 | |
| stevemar | bknudson, it would be an invalid map, so it wouldn't hit the processor | 21:47 |
| *** jcooley_ has quit IRC | 21:48 | |
| bknudson | stevemar: well, only have to test valid mappings, but it should be a little more obvious what it's testing... | 21:48 |
| bknudson | it should be obvious that we've got a test for "any_one_of" | 21:48 |
| dstanek | morganfainberg: you around? | 21:48 |
| bknudson | and a test for "not_any_of", and a test for direct maps. | 21:48 |
| bknudson | and a test for regex | 21:48 |
| *** d0ugal has quit IRC | 21:49 | |
| *** jcooley_ has joined #openstack-dev | 21:49 | |
| *** coasterz has joined #openstack-dev | 21:49 | |
| stevemar | bknudson, ahhhh, okay, I can rename the tests, if that's better, and make sure the comments describe the test | 21:49 |
| bknudson | stevemar: that would be great | 21:49 |
| *** NikitaKonovalov is now known as NikitaKonovalov_ | 21:50 | |
| *** romcheg1 has joined #openstack-dev | 21:51 | |
| *** romcheg_ltp has quit IRC | 21:51 | |
| *** Sukhdev has joined #openstack-dev | 21:51 | |
| *** jasondotstar has quit IRC | 21:51 | |
| *** Longgeek has joined #openstack-dev | 21:51 | |
| *** tjones has quit IRC | 21:51 | |
| *** Longgeek has quit IRC | 21:51 | |
| *** david-lyle has quit IRC | 21:51 | |
| *** Longgeek has joined #openstack-dev | 21:52 | |
| morganfainberg | dstanek, hi | 21:52 |
| *** mfer has quit IRC | 21:52 | |
| morganfainberg | dstanek, was getting coffee and having ad-hoc meeting at the coffee shop ... i know /slacker | 21:52 |
| *** Longgeek has quit IRC | 21:52 | |
| dstanek | morganfainberg: :-) looking at https://review.openstack.org/#/c/71683/2/keystone/token/backends/memcache.py | 21:52 |
| dstanek | on line 115 would you be getting data in the old format? | 21:53 |
| morganfainberg | dstanek, correct | 21:53 |
| *** prad has joined #openstack-dev | 21:53 | |
| *** dvarga has quit IRC | 21:54 | |
| *** mrodden has joined #openstack-dev | 21:54 | |
| *** Longgeek has joined #openstack-dev | 21:54 | |
| morganfainberg | dstanek, or it might be the key has expired and is empty | 21:54 |
| *** Longgeek has quit IRC | 21:54 | |
| morganfainberg | dstanek, oh, no no nvm None is checked on line 114 | 21:54 |
| morganfainberg | dstanek, your assumption is correct | 21:55 |
| *** Longgeek has joined #openstack-dev | 21:56 | |
| morganfainberg | dstanek, however, the core of the issue is that asking memcached a bunch for tokens locks up keystone, so we kind of need to start from scratch for that list. | 21:56 |
| dstanek | morganfainberg: do we need to convert that to the new format or is it ok if is gets lost? | 21:56 |
| morganfainberg | dstanek, ideally...memcached should be flushed before this code goes live | 21:56 |
| *** Longgeek has quit IRC | 21:57 | |
| morganfainberg | dstanek, but we don't know if tokens are valid or not, we can't ask w/o possibly locking up keystone truying to ask for the data | 21:57 |
| *** jasondotstar has joined #openstack-dev | 21:57 | |
| *** clayb has joined #openstack-dev | 21:57 | |
| morganfainberg | dstanek, i think the only real solution is to clear the list and force the user(s) to get new tokens. | 21:57 |
| *** sarob has joined #openstack-dev | 21:58 | |
| morganfainberg | dstanek, it's an unfortunate design choice i made when i originally developed this code. | 21:58 |
| *** Longgeek has joined #openstack-dev | 21:58 | |
| dstanek | morganfainberg: so when this code is deployed all token are expired? | 21:58 |
| morganfainberg | dstanek, not... exactly. | 21:58 |
| *** Longgeek has quit IRC | 21:59 | |
| morganfainberg | dstanek, tokens are no longer tracked. | 21:59 |
| morganfainberg | dstanek, i guess this means there is a <token_TTL> window of possible token abuse | 21:59 |
| *** Longgeek has joined #openstack-dev | 21:59 | |
| morganfainberg | let me look at something real quickly | 21:59 |
| bknudson | tokens that were valid before can't be validated via keystone... UUID tokens | 21:59 |
| morganfainberg | bknudson, correct, but PKI | 22:00 |
| *** Longgeek has quit IRC | 22:00 | |
| bknudson | auth_token won't care about the pki tokens | 22:00 |
| *** anniec has quit IRC | 22:00 | |
| bknudson | but then keystone can't invalidate them (due to role change or whatever?) | 22:00 |
| morganfainberg | bknudson, you can't revoke PKI tokens | 22:00 |
| morganfainberg | bknudson, yeah | 22:00 |
| morganfainberg | bknudson, that is the issue | 22:00 |
| *** Longgeek has joined #openstack-dev | 22:01 | |
| *** dspano has quit IRC | 22:01 | |
| morganfainberg | bknudson, for uuid tokens, i expected it to impact the validity of them. | 22:02 |
| morganfainberg | bknudson, and that is covered in the commit message - a side effect of the fix really being required. | 22:02 |
| *** Longgeek has quit IRC | 22:02 | |
| morganfainberg | bknudson, not sure how to address the password-change/role-change/etc revocation of tokens for the token-ttl window after this is deployed | 22:03 |
| morganfainberg | i... guess we could just use current token-ttl and keep the elements of this list around longer than before? | 22:03 |
| *** Longgeek has joined #openstack-dev | 22:03 | |
| *** marcoemorais has quit IRC | 22:03 | |
| morganfainberg | bknudson, read in the list like we used to, hard-set token TTL to now+token_TTL time from conf, and then loop? | 22:03 |
| *** Longgeek has quit IRC | 22:04 | |
| *** marcoemorais has joined #openstack-dev | 22:04 | |
| morganfainberg | it only means the user-index would hold tokens that aren't valid anymore for longer. it's expected some tokens will be invalid and in that list | 22:04 |
| *** marcoemorais has quit IRC | 22:04 | |
| morganfainberg | dstanek, ^ | 22:04 |
| *** READ10 has quit IRC | 22:05 | |
| *** Longgeek has joined #openstack-dev | 22:05 | |
| *** marcoemorais has joined #openstack-dev | 22:06 | |
| bknudson | morganfainberg: I think that sounds reasonable... essentially fake out the tokens in the existing list with an expiration. | 22:06 |
| *** Longgeek has quit IRC | 22:06 | |
| *** vuil has joined #openstack-dev | 22:06 | |
| dstanek | morganfainberg, bknudson: yes, that is what i was thinking | 22:06 |
| morganfainberg | ok, that... shouldn't be too bad. it has a potential gap in that if someone changed the TTL in the conf, but -- really I don't think we can address every edge-case | 22:06 |
| *** Longgeek has joined #openstack-dev | 22:06 | |
| *** thuc has quit IRC | 22:07 | |
| *** Longgeek has quit IRC | 22:07 | |
| *** jdob_ has quit IRC | 22:07 | |
| morganfainberg | ok, i'll add some timedelta magic code in there. | 22:07 |
| morganfainberg | s/magic// | 22:07 |
| *** thuc has joined #openstack-dev | 22:07 | |
| morganfainberg | and i'll port that to the havana version as well. | 22:07 |
| morganfainberg | any other comments before i post a patchset? [don't want to cause issues with reviews/extra patches since this is important fix] | 22:08 |
| *** Longgeek has joined #openstack-dev | 22:09 | |
| *** jasondotstar has quit IRC | 22:09 | |
| *** Longgeek has quit IRC | 22:09 | |
| *** Longgeek has joined #openstack-dev | 22:09 | |
| *** Longgeek has quit IRC | 22:10 | |
| dstanek | morganfainberg: the only other thing that struck me was line 172 on https://review.openstack.org/#/c/71683/2/keystone/tests/test_backend_memcache.py | 22:11 |
| *** thuc has quit IRC | 22:11 | |
| *** thuc has joined #openstack-dev | 22:12 | |
| *** Longgeek has joined #openstack-dev | 22:12 | |
| dstanek | i think it would be useful to call out that the next few lines are really to expire one of the tokens | 22:12 |
| *** abhirc has joined #openstack-dev | 22:12 | |
| *** Longgeek has quit IRC | 22:13 | |
| dstanek | or simplify it to just expire one of them and not look for the one labeled as expired. i don't think you look to make sure the expired one was actually expired anyway | 22:13 |
| *** tjones has joined #openstack-dev | 22:14 | |
| *** lparth has joined #openstack-dev | 22:15 | |
| *** Longgeek has joined #openstack-dev | 22:15 | |
| *** Longgeek has quit IRC | 22:15 | |
| bknudson | morganfainberg: I posted a couple | 22:15 |
| bknudson | dstanek: that was my comment. | 22:16 |
| *** Longgeek has joined #openstack-dev | 22:16 | |
| bknudson | we think alike | 22:16 |
| *** Longgeek has quit IRC | 22:16 | |
| *** Longgeek has joined #openstack-dev | 22:17 | |
| *** sballe has quit IRC | 22:18 | |
| *** Longgeek has quit IRC | 22:18 | |
| *** tjones has quit IRC | 22:18 | |
| *** Longgeek has joined #openstack-dev | 22:18 | |
| *** bswartz has quit IRC | 22:19 | |
| *** carl_baldwin has quit IRC | 22:21 | |
| *** Longgeek has quit IRC | 22:21 | |
| *** bknudson has quit IRC | 22:21 | |
| *** Longgeek has joined #openstack-dev | 22:22 | |
| *** peristeri has quit IRC | 22:23 | |
| *** otherwiseguy has quit IRC | 22:23 | |
| *** Longgeek has quit IRC | 22:24 | |
| *** julienvey has quit IRC | 22:24 | |
| dstanek | :-) | 22:24 |
| *** Longgeek has joined #openstack-dev | 22:24 | |
| *** rtheis has quit IRC | 22:26 | |
| *** arnaud has quit IRC | 22:26 | |
| *** _cjones__ has joined #openstack-dev | 22:27 | |
| *** melwitt has quit IRC | 22:27 | |
| *** csd has joined #openstack-dev | 22:27 | |
| *** abhirc has quit IRC | 22:27 | |
| *** Longgeek has quit IRC | 22:27 | |
| *** vladikr has quit IRC | 22:27 | |
| *** chris_johnson has quit IRC | 22:28 | |
| *** epim has quit IRC | 22:28 | |
| *** Longgeek has joined #openstack-dev | 22:28 | |
| *** vuil has quit IRC | 22:28 | |
| *** godara has joined #openstack-dev | 22:29 | |
| *** prad has quit IRC | 22:29 | |
| *** tjones has joined #openstack-dev | 22:29 | |
| *** marcoemorais has quit IRC | 22:30 | |
| *** _cjones_ has quit IRC | 22:30 | |
| *** marcoemorais has joined #openstack-dev | 22:30 | |
| *** jcooley_ has quit IRC | 22:31 | |
| *** Longgeek has quit IRC | 22:31 | |
| *** Longgeek has joined #openstack-dev | 22:31 | |
| *** arnaud has joined #openstack-dev | 22:31 | |
| *** tjones has quit IRC | 22:32 | |
| *** tjones has joined #openstack-dev | 22:32 | |
| *** epim has joined #openstack-dev | 22:32 | |
| *** melwitt has joined #openstack-dev | 22:32 | |
| *** _cjones__ has quit IRC | 22:34 | |
| *** Longgeek has quit IRC | 22:34 | |
| *** BLZbubba has quit IRC | 22:34 | |
| *** BLZbubba has joined #openstack-dev | 22:34 | |
| *** Longgeek has joined #openstack-dev | 22:35 | |
| *** sushils has joined #openstack-dev | 22:35 | |
| stevemar | i think i went overboard on docstrings :\ | 22:37 |
| *** sarob has quit IRC | 22:37 | |
| *** pablosan has joined #openstack-dev | 22:37 | |
| annegentle | stevemar: unpossible | 22:37 |
| stevemar | annegentle, i think the docstring block is longer than the code now | 22:37 |
| *** Longgeek has quit IRC | 22:37 | |
| dstanek | stevemar: if they are helpful then that's a good thing | 22:38 |
| *** carl_baldwin has joined #openstack-dev | 22:38 | |
| annegentle | stevemar: :) | 22:38 |
| *** anniec has joined #openstack-dev | 22:38 | |
| *** Longgeek has joined #openstack-dev | 22:38 | |
| morganfainberg | dstanek, thanks | 22:38 |
| *** Longgeek has quit IRC | 22:38 | |
| morganfainberg | dstanek, i'll look at both your comments (in IRC) and brant's | 22:39 |
| *** sarob has joined #openstack-dev | 22:39 | |
| *** Longgeek has joined #openstack-dev | 22:39 | |
| dstanek | morganfainberg: np | 22:39 |
| *** Longgeek has quit IRC | 22:39 | |
| *** Longgeek has joined #openstack-dev | 22:40 | |
| *** anniec has quit IRC | 22:41 | |
| *** Mandeep has joined #openstack-dev | 22:42 | |
| *** Longgeek has quit IRC | 22:43 | |
| *** sarob has quit IRC | 22:43 | |
| *** bswartz has joined #openstack-dev | 22:43 | |
| *** anniec has joined #openstack-dev | 22:43 | |
| *** Longgeek has joined #openstack-dev | 22:44 | |
| Mandeep | help | 22:44 |
| *** pmathews has quit IRC | 22:44 | |
| *** colinmcnamara has joined #openstack-dev | 22:44 | |
| *** ewindisch has quit IRC | 22:44 | |
| *** bknudson has joined #openstack-dev | 22:46 | |
| *** mriedem has quit IRC | 22:46 | |
| *** Longgeek has quit IRC | 22:46 | |
| *** Longgeek has joined #openstack-dev | 22:47 | |
| *** Longgeek has quit IRC | 22:48 | |
| *** colinmcnamara has quit IRC | 22:49 | |
| *** colinmcnamara has joined #openstack-dev | 22:49 | |
| ayoung | dolphm, we can drop support for the MII format once we have the {cmsz} format | 22:50 |
| ayoung | dstanek, sorry, missed your review | 22:53 |
| ayoung | dolphm, I'm OK with dropping the MII format if you are | 22:53 |
| ayoung | but this code will have to deal with both. | 22:53 |
| ayoung | at least until we can get the server to stop producing the MII format | 22:54 |
| *** vkmc has joined #openstack-dev | 22:54 | |
| *** nati_ueno has quit IRC | 22:54 | |
| *** kevinconway has quit IRC | 22:55 | |
| *** nati_ueno has joined #openstack-dev | 22:55 | |
| *** ewindisch has joined #openstack-dev | 22:56 | |
| dolphm | ayoung: drop support for generating MII tokens - that's all. still support reading them in auth_token (the code in auth_token looks great afaict) | 22:56 |
| dolphm | other than i would swap the default value of DER vs PEM | 22:56 |
| dolphm | in whatever kwarg that was called (inform/outform?) | 22:56 |
| *** nkinder has joined #openstack-dev | 22:57 | |
| ayoung | dolphm, I'd like to do it in two stages. I think dropping MII should be second. If we merge the change that lets server use the client lib, and we drop support for MII at the same time, we'll be unable to merge. We need to remove the change from the server side first | 22:57 |
| ayoung | and I really don;t want to duplicate this code onto both sides | 22:57 |
| dolphm | ayoung: *thinking* | 22:58 |
| *** lcheng_ has joined #openstack-dev | 22:58 | |
| ayoung | I hear you on the DER vs PEM thing, but again, I was treating it like a public API, since the server is calling the cloned version of this code in keystone/common | 22:58 |
| ayoung | I can put in deprecations for now | 22:58 |
| dolphm | fair enough ^ | 22:59 |
| ayoung | personally, I love the idea of dropping MII | 22:59 |
| *** edmund has quit IRC | 22:59 | |
| dstanek | morganfainberg: let me know when you're ready with the memcached review and i'll get on it right away | 22:59 |
| ayoung | let me address dstanek 's comments, I didn;t even realize that review was in there... | 22:59 |
| morganfainberg | dstanek, working on the fixes now | 22:59 |
| dolphm | ayoung: i'm not understanding why you need to do the client side in two patches? (that means two 0.x.0 releases?) | 22:59 |
| dstanek | morganfainberg: take your time - i have all night :-) | 22:59 |
| *** jnoller_ has joined #openstack-dev | 23:01 | |
| *** rwsu has quit IRC | 23:01 | |
| ayoung | dolphm, yeah...I don't like it either. Let me think it through and see if there is a way we can do it without breaking the server. | 23:01 |
| *** jnoller has quit IRC | 23:01 | |
| *** byeager has joined #openstack-dev | 23:01 | |
| *** jcooley_ has joined #openstack-dev | 23:01 | |
| ayoung | dolphm, OK, so, in order to do that, I would have to do one of two things. Either clone the change to the server, then change the server to use the library. Or, not clone the change to the server, and merge the change to use the library with the change to not produce MII tokens | 23:03 |
| *** rwsu has joined #openstack-dev | 23:03 | |
| dolphm | ayoung: the server isn't consuming the client's cms module yet, right? | 23:03 |
| ayoung | dolphm, yeah, not yet | 23:03 |
| dolphm | ayoung: so the token generation code is unused there? | 23:03 |
| ayoung | due to what dstanek found about the check for Popen | 23:03 |
| ayoung | token generation code in the server is performed in keystone/common/cms.py | 23:04 |
| ayoung | signing that is | 23:04 |
| *** _cjones_ has joined #openstack-dev | 23:04 | |
| *** jobewan has quit IRC | 23:04 | |
| *** jsavak has quit IRC | 23:04 | |
| *** _cjones_ has quit IRC | 23:05 | |
| dolphm | ayoung: there's two steps required, right? A) client release supports generating compressed tokens, B) service is updated to use client's cms module | 23:05 |
| *** _cjones_ has joined #openstack-dev | 23:05 | |
| dolphm | i don't see how dropping support for generating MII in step A would break the server if it's not even used | 23:05 |
| *** rwsu has quit IRC | 23:06 | |
| *** jcooley_ has quit IRC | 23:06 | |
| ayoung | dolphm, OK, if the client drops support for generating MII tokens, I can do it all cleanly with 3 server patches. | 23:07 |
| ayoung | 1: move the generation of MII tokens out of common/cms.py into a temp file | 23:07 |
| ayoung | 2: switch to using the client lib | 23:07 |
| ayoung | 3: drop the MII tokens | 23:07 |
| ayoung | so, I can drop support for the MII tokens out of the client lib. I'll do that this patch. | 23:08 |
| *** otherwiseguy has joined #openstack-dev | 23:08 | |
| *** jdennis has quit IRC | 23:08 | |
| ayoung | we can hold up on approving patch 1 above until 2 and 3 are approved, and it will pass gerrit as one virtual transaction | 23:08 |
| *** Mandeep has quit IRC | 23:08 | |
| *** RajeshMohan has quit IRC | 23:09 | |
| *** jmontemayor has quit IRC | 23:10 | |
| *** burt1 has quit IRC | 23:10 | |
| *** RajeshMohan has joined #openstack-dev | 23:10 | |
| dolphm | ayoung: what does step 3 entail? | 23:10 |
| *** browne has quit IRC | 23:11 | |
| *** ewindisch has quit IRC | 23:11 | |
| dolphm | ayoung: you won't be able to pass a jenkins check in step 2 until the client sees a release in step 1 though | 23:11 |
| *** galstrom is now known as galstrom_zzz | 23:12 | |
| *** joesavak has joined #openstack-dev | 23:14 | |
| *** jgrimm has quit IRC | 23:15 | |
| *** tdruiva has quit IRC | 23:16 | |
| ayoung | dolphm, right, so in step 1, I move all deprecated/removed functions out of keystone/common/cms.py and to a temp file. Step 2 is to drop keystone/common/cms.py | 23:16 |
| ayoung | step 3 entails changing over all of the MII specific code to code that checks for and generats cmsz format, and drops the temp file. | 23:17 |
| ayoung | unit tests run to 100% on each step | 23:17 |
| dolphm | ayoung: keystone/common/cms isn't public api -- don't worry about managing the transition like that | 23:17 |
| *** marc_ has joined #openstack-dev | 23:18 | |
| ayoung | dolphm, well, I can do it all in one commit, but it will be easier to review in the 3 steps above. | 23:18 |
| *** marc_ is now known as Guest45108 | 23:18 | |
| ayoung | I need to go be a dad for a while. | 23:18 |
| *** troytoman-away is now known as troytoman | 23:18 | |
| dolphm | ayoung: /salute | 23:18 |
| *** ayoung is now known as ayoung-ZzZz | 23:18 | |
| *** markmcclain has joined #openstack-dev | 23:18 | |
| *** rwsu has joined #openstack-dev | 23:19 | |
| *** markmcclain has quit IRC | 23:20 | |
| *** markmcclain has joined #openstack-dev | 23:21 | |
| *** enikanorov has quit IRC | 23:21 | |
| *** byeager has quit IRC | 23:21 | |
| *** FunnyLookinHat has quit IRC | 23:21 | |
| *** IanGovett has quit IRC | 23:21 | |
| *** thuc has quit IRC | 23:21 | |
| *** jasondotstar has joined #openstack-dev | 23:22 | |
| *** thuc has joined #openstack-dev | 23:22 | |
| *** enikanorov has joined #openstack-dev | 23:24 | |
| *** doug_shelley66 has quit IRC | 23:24 | |
| *** doug_shelley66 has joined #openstack-dev | 23:25 | |
| *** tdruiva has joined #openstack-dev | 23:25 | |
| *** jnoller_ is now known as jnoller | 23:25 | |
| *** thuc has quit IRC | 23:26 | |
| *** radsy has joined #openstack-dev | 23:27 | |
| dolphm | dhellmann: what's the impact on all the projects that have been added to bug 1277507 ? | 23:29 |
| uvirtbot | Launchpad bug 1277507 in python-keystoneclient ""ImportError: No module named passlib.hash"; HTTP error 403 while getting ipaddr from googledrive.com" [Undecided,Invalid] https://launchpad.net/bugs/1277507 | 23:29 |
| dhellmann | dolphm: at this point ipaddr is available on pypi and so things should be working again | 23:29 |
| *** Guest45108 has quit IRC | 23:29 | |
| dhellmann | dolphm: I've been working on some internal stuff, though, so maybe something new is going on I don't know about? | 23:30 |
| dolphm | dhellmann: ++, do we need to bump requirements or something to avoid the issue recurring? | 23:30 |
| dhellmann | dolphm: let me look at what requirements says now | 23:30 |
| *** mst89 has joined #openstack-dev | 23:30 | |
| dhellmann | if there's only one version available, the mirror sync should pick that one up | 23:30 |
| *** thuc has joined #openstack-dev | 23:30 | |
| dhellmann | dolphm: ipaddr isn't in our requirements, I guess it's a 2nd tier requirement | 23:31 |
| dolphm | dhellmann: pretty sure WSME uses it? | 23:31 |
| *** thuc_ has joined #openstack-dev | 23:31 | |
| *** achampion has quit IRC | 23:31 | |
| *** anniec has quit IRC | 23:31 | |
| dhellmann | dolphm: yeah | 23:31 |
| dhellmann | I thought we had a direct dependency, too | 23:31 |
| dolphm | dhellmann: i thought we did as well, actually | 23:32 |
| *** ijw has quit IRC | 23:32 | |
| *** thuc_ has quit IRC | 23:32 | |
| jnoller | dhellmann: did he upload it to pypi | 23:32 |
| jnoller | ? | 23:32 |
| *** thuc has quit IRC | 23:32 | |
| dolphm | jnoller: yes | 23:32 |
| *** thuc has joined #openstack-dev | 23:32 | |
| *** joesavak has quit IRC | 23:32 | |
| *** arnaud__ has quit IRC | 23:33 | |
| *** arnaud has quit IRC | 23:33 | |
| *** morazi has quit IRC | 23:33 | |
| *** lbragstad has quit IRC | 23:34 | |
| *** adnan has quit IRC | 23:34 | |
| *** epim has quit IRC | 23:34 | |
| jnoller | Woot | 23:34 |
| jnoller | he didn't reply to my email so :( | 23:34 |
| dhellmann | jnoller: he replied on their dev list | 23:35 |
| jnoller | this is why I can't have nice things | 23:35 |
| * dhellmann is getting tired of subscribing to dev lists to communicate with package authors | 23:35 | |
| *** vuil has joined #openstack-dev | 23:35 | |
| *** arnaud has joined #openstack-dev | 23:36 | |
| *** arnaud__ has joined #openstack-dev | 23:36 | |
| *** thuc has quit IRC | 23:37 | |
| *** vuil has quit IRC | 23:38 | |
| *** vuil has joined #openstack-dev | 23:38 | |
| *** pablosan has quit IRC | 23:39 | |
| *** neelashah has quit IRC | 23:40 | |
| *** anniec has joined #openstack-dev | 23:40 | |
| *** thuc has joined #openstack-dev | 23:40 | |
| *** aaronjamesford has quit IRC | 23:42 | |
| *** markmcclain has quit IRC | 23:43 | |
| *** thuc has quit IRC | 23:44 | |
| *** e0ne has quit IRC | 23:44 | |
| *** thuc has joined #openstack-dev | 23:44 | |
| *** tjones1 has joined #openstack-dev | 23:45 | |
| *** denis_makogon_ has quit IRC | 23:45 | |
| *** tjones has quit IRC | 23:46 | |
| *** e0ne has joined #openstack-dev | 23:46 | |
| *** tjones1 has quit IRC | 23:48 | |
| *** tjones has joined #openstack-dev | 23:48 | |
| *** browne has joined #openstack-dev | 23:49 | |
| *** Sukhdev has quit IRC | 23:51 | |
| *** tjones has quit IRC | 23:53 | |
| *** jnoller has quit IRC | 23:54 | |
| *** anniec has quit IRC | 23:56 | |
| *** vuil has quit IRC | 23:57 | |
| *** abhirc has joined #openstack-dev | 23:58 | |
| *** MaxV has quit IRC | 23:59 | |
| *** jkoelker has quit IRC | 23:59 | |
| *** jkoelker has joined #openstack-dev | 23:59 | |
| *** clu_ has quit IRC | 23:59 | |
| *** vartom1111111117 has quit IRC | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
