| *** reldan has joined #openstack | 00:00 | |
| *** pvo has quit IRC | 00:00 | |
| devcamcar | jk0, kpepple: review is up: https://code.launchpad.net/~ken-pepple/nova/dynamicinstancetypes/+merge/50257 | 00:00 |
|---|---|---|
| devcamcar | back in a few | 00:01 |
| *** littleidea has quit IRC | 00:01 | |
| jk0 | cool | 00:01 |
| jk0 | kpepple: we might get this yet tonight :) | 00:01 |
| *** rds__ has quit IRC | 00:01 | |
| kpepple | jk0: maybe | 00:02 |
| kpepple | jk0: how do you want to handle the style issues ? | 00:02 |
| jk0 | I wasn't aware of any written guidelines on stuff like that, but I suppose it wouldn't hurt to fix them up since it was requested in the review | 00:03 |
| jk0 | (I personally prefer it the way it is) | 00:03 |
| *** joearnold has quit IRC | 00:05 | |
| kpepple | jk0: okay, i'm updating nova-manage | 00:05 |
| *** mray has quit IRC | 00:06 | |
| *** vvuksan has quit IRC | 00:12 | |
| kpepple | jk0: i'm almost through all of them ... just running tests now. | 00:17 |
| *** johnpur has quit IRC | 00:21 | |
| *** grapex has left #openstack | 00:28 | |
| devcamcar | jk0: style nits were just to keep codebase consistent | 00:29 |
| jk0 | I hear ya | 00:30 |
| kpepple | devcamcar: on your last review note (954), don't i need the context as for the decorator ? | 00:34 |
| devcamcar | kpepple: shouldn't, its just a positional arg | 00:35 |
| *** mustfg has quit IRC | 00:39 | |
| *** adiantum has quit IRC | 00:42 | |
| *** dubsquared has quit IRC | 00:42 | |
| *** adiantum has joined #openstack | 00:47 | |
| kpepple | devcamcar: on the "Why are all of these removed? There are a lot more." in regards to the docs -- i spoke with anne about this yesterday and she said it can be either way (they are generated docs) ... i'll put them back in | 00:48 |
| devcamcar | kpepple: if its stuff that is autogenerated it probably doesn't belong there. i guess do whatever anne thinks is best there :) | 00:51 |
| devcamcar | kpepple: though if we're going to remove stuff like that its better to do it in its own branch | 00:51 |
| kpepple | devcamcar: do you have a more preferred way of raising DB errors generated through the API ? or is just logging the DB error fine ... | 01:01 |
| *** pvo has joined #openstack | 01:01 | |
| jk0 | kpepple: what lines are you referring to? the DBException (or whatever it's called)? | 01:02 |
| kpepple | jk0: the "exception.NotFound" from instance_types.py | 01:02 |
| devcamcar | kpepple: not sure, let me see if i can find an example | 01:03 |
| openstackhudson | Project nova build #593: SUCCESS in 1 min 48 sec: http://hudson.openstack.org/job/nova/593/ | 01:03 |
| openstackhudson | Tarmac: Fixes Bug #715424: nova-manage : create network crashes when subnet range provided is not enough , if the network range cannot fit the parameters passed, a ValueError is raised | 01:03 |
| uvirtbot | Launchpad bug 715424 in nova "nova-manage : create network crashes when subnet range provided is not enough" [Low,Fix committed] https://launchpad.net/bugs/715424 | 01:03 |
| jk0 | ah | 01:03 |
| jk0 | hm | 01:03 |
| jk0 | I think we could get by logging the DB error (if that's where the real error lies) | 01:04 |
| jk0 | unless devcamcar is referring to something else | 01:04 |
| devcamcar | jk0: yea | 01:04 |
| *** pvo has quit IRC | 01:05 | |
| kpepple | devcamcar , jk0 : ok, setting up a nova.instance_types log and i'll LOG.exception them | 01:05 |
| jk0 | ah yeah, .exception is the way to go | 01:05 |
| jk0 | especially if we don't know where the exception is going to come from | 01:06 |
| *** clauden_ has quit IRC | 01:06 | |
| jk0 | (just found what devcamcar was referring to) | 01:06 |
| devcamcar | word | 01:08 |
| openstackhudson | Project nova build #594: SUCCESS in 1 min 44 sec: http://hudson.openstack.org/job/nova/594/ | 01:08 |
| openstackhudson | Tarmac: Very simple change checking for < 0 values in "limit" and "offset" GET parameters. If either are negative, raise a HTTPBadRequest exception. Relevant tests included. | 01:08 |
| *** adiantum has quit IRC | 01:12 | |
| *** reldan has quit IRC | 01:14 | |
| *** MotoMilind has quit IRC | 01:14 | |
| nelson | woot | 01:16 |
| kpepple | jk0 , devcamcar : just pushed revno 696 which i believe addresses all the code review concerns | 01:17 |
| jk0 | excellent | 01:17 |
| *** adiantum has joined #openstack | 01:17 | |
| devcamcar | nice | 01:18 |
| *** sandywalsh has quit IRC | 01:18 | |
| *** perestrelka has quit IRC | 01:19 | |
| *** perestrelka has joined #openstack | 01:19 | |
| *** burris has quit IRC | 01:19 | |
| *** gregp76_ has joined #openstack | 01:24 | |
| *** littleidea has joined #openstack | 01:25 | |
| jk0 | kpepple / devcamcar shall we get this merged? :) | 01:26 |
| *** cascone has joined #openstack | 01:28 | |
| devcamcar | 34+ instance_types.create(name, memory, vcpus, local_gb, | 01:29 |
| devcamcar | 35+ flavorid, swap, rxtx_quota, rxtx_cap | 01:29 |
| devcamcar | you can put that on multi lines if you want | 01:29 |
| devcamcar | its just for the defs | 01:29 |
| devcamcar | so | 01:29 |
| devcamcar | .create(name, | 01:30 |
| devcamcar | memory, | 01:30 |
| devcamcar | etc | 01:30 |
| devcamcar | is fine | 01:30 |
| jk0 | oh, *now* you like my method! | 01:30 |
| jk0 | :P jjk | 01:30 |
| devcamcar | hah | 01:33 |
| devcamcar | kpepple: 817+ except: | 01:33 |
| devcamcar | pep8 probably going to complain about that | 01:33 |
| kpepple | devcamcar: surprisingly it didn't (i have the pep8 bzr-commit hook) | 01:34 |
| devcamcar | cool | 01:34 |
| kpepple | devcamcar: but noted ... let me change that | 01:34 |
| devcamcar | kpepple: gave it the lgtm | 01:36 |
| devcamcar | nice work! | 01:36 |
| jk0 | \o/ | 01:36 |
| jk0 | thanks devcamcar | 01:37 |
| kpepple | devcamcar: thx for all the help on this | 01:37 |
| kpepple | jk0: let me push this one last change in | 01:38 |
| jk0 | werd | 01:38 |
| devcamcar | i'm off, later gang | 01:38 |
| jk0 | night night | 01:38 |
| *** perestrelka has quit IRC | 01:39 | |
| *** pyros_xx has quit IRC | 01:39 | |
| *** dweimer has quit IRC | 01:39 | |
| *** maplebed has quit IRC | 01:39 | |
| *** dysinger has quit IRC | 01:39 | |
| *** eday has quit IRC | 01:39 | |
| *** cw has quit IRC | 01:39 | |
| jk0 | kpepple: I'll approve it when you push that thru | 01:39 |
| *** dweimer has joined #openstack | 01:39 | |
| *** adiantum has quit IRC | 01:41 | |
| kpepple | jk0: just pushed 697 which should be it | 01:41 |
| * jk0 waits for the diff update | 01:41 | |
| *** perestrelka has joined #openstack | 01:43 | |
| *** pyros_xx has joined #openstack | 01:43 | |
| *** maplebed has joined #openstack | 01:43 | |
| *** dysinger has joined #openstack | 01:43 | |
| *** eday has joined #openstack | 01:43 | |
| *** cw has joined #openstack | 01:43 | |
| *** niven.freenode.net sets mode: +v eday | 01:43 | |
| *** pyros_xx has quit IRC | 01:43 | |
| jk0 | kpepple: approved. should merge whenever LP gets around to it :) | 01:44 |
| kpepple | jk0: very cool. thank for all the help on this. | 01:44 |
| * kpepple wanders off in search of frosty beverages | 01:45 | |
| jk0 | no, thank YOU (you saved me a lot of work) :) | 01:45 |
| jk0 | yeah, it is beer time now | 01:45 |
| jk0 | l8r man | 01:45 |
| *** perestrelka has quit IRC | 01:45 | |
| *** maplebed has quit IRC | 01:45 | |
| *** dysinger has quit IRC | 01:45 | |
| *** eday has quit IRC | 01:45 | |
| *** cw has quit IRC | 01:45 | |
| *** dovetaildan has joined #openstack | 01:45 | |
| kpepple | jk0: later | 01:48 |
| jk0 | kpepple: wait | 01:48 |
| jk0 | kpepple: tests failed on hudson | 01:49 |
| kpepple | jk0: you got a link ... tests run fine on my machine | 01:49 |
| jk0 | kpepple: oopsies -- rename your migration file to 008 | 01:49 |
| jk0 | I added 007 this evening :o | 01:49 |
| jk0 | sorry, should have caught that | 01:49 |
| * kpepple thinks migration files are going to be the death of him | 01:49 | |
| jk0 | haha | 01:49 |
| jk0 | didn't we start out on like 003? | 01:49 |
| kpepple | jk0: yes. let me merge trunk and mv the file over. | 01:50 |
| *** littleidea has quit IRC | 01:50 | |
| *** perestrelka has joined #openstack | 01:51 | |
| *** maplebed has joined #openstack | 01:51 | |
| *** dysinger has joined #openstack | 01:51 | |
| *** eday has joined #openstack | 01:51 | |
| *** cw has joined #openstack | 01:51 | |
| *** niven.freenode.net sets mode: +v eday | 01:51 | |
| *** adiantum has joined #openstack | 01:52 | |
| jk0 | well, at least it's an easy fix :) | 01:53 |
| kpepple | jk0: yeah, just running the tests | 01:53 |
| kpepple | jk0: okay, merged trunk, bzr mv migration file and re-pushed at revno 699 | 01:54 |
| *** gregp76_ has quit IRC | 01:54 | |
| jk0 | werd, one sec | 01:54 |
| jk0 | waiting on LP of course | 01:55 |
| jk0 | *caugh* wish we used github *caugh* | 01:55 |
| * jk0 sweeps dead horse back under matt for now | 01:56 | |
| *** dendrobates is now known as dendro-afk | 01:56 | |
| jk0 | kpepple: approving. standby | 01:58 |
| *** burris has joined #openstack | 02:02 | |
| jk0 | kpepple: merged. go on about your business ;) | 02:07 |
| *** Ryan_Lane has joined #openstack | 02:07 | |
| openstackhudson | Project nova build #595: SUCCESS in 1 min 47 sec: http://hudson.openstack.org/job/nova/595/ | 02:07 |
| openstackhudson | Tarmac: merges dynamic instance types blueprint (http://wiki.openstack.org/ConfigureInstanceTypesDynamically) and bundles blueprint (https://blueprints.launchpad.net/nova/+spec/flavors) | 02:08 |
| kpepple | jk0: excellent. thx. | 02:08 |
| jk0 | you betcha~ | 02:09 |
| *** Ryan_Lane has quit IRC | 02:13 | |
| *** adjohn has quit IRC | 02:20 | |
| *** adiantum has quit IRC | 02:24 | |
| *** adiantum has joined #openstack | 02:25 | |
| *** adiantum has quit IRC | 02:31 | |
| *** adiantum has joined #openstack | 02:31 | |
| *** sandywalsh has joined #openstack | 02:34 | |
| *** mray has joined #openstack | 02:34 | |
| *** adiantum has quit IRC | 02:38 | |
| *** adiantum has joined #openstack | 02:50 | |
| *** MarkAtwood has joined #openstack | 02:50 | |
| *** MarkusT has quit IRC | 03:05 | |
| *** adiantum has quit IRC | 03:09 | |
| *** syah has quit IRC | 03:11 | |
| *** syah has joined #openstack | 03:13 | |
| *** adiantum has joined #openstack | 03:16 | |
| *** MarkusT has joined #openstack | 03:19 | |
| *** mdomsch has joined #openstack | 03:19 | |
| *** pvo has joined #openstack | 03:20 | |
| *** lamar has joined #openstack | 03:22 | |
| *** dendro-afk is now known as dendrobates | 03:23 | |
| *** mray has quit IRC | 03:31 | |
| *** mahadev has quit IRC | 03:35 | |
| *** RJD22 is now known as RJD22|away | 03:37 | |
| *** RJD22|away is now known as RJD22 | 03:48 | |
| *** rchavik has joined #openstack | 03:48 | |
| *** mahadev has joined #openstack | 03:52 | |
| *** adiantum has quit IRC | 03:53 | |
| *** RJD22 is now known as RJD22|away | 03:53 | |
| *** mahadev has quit IRC | 03:56 | |
| *** adiantum has joined #openstack | 03:56 | |
| *** thielmann_ has joined #openstack | 03:59 | |
| *** MarkAtwood has quit IRC | 03:59 | |
| *** kainam has quit IRC | 04:00 | |
| *** arcane has joined #openstack | 04:00 | |
| *** MarkusT has quit IRC | 04:02 | |
| *** mahadev has joined #openstack | 04:03 | |
| *** RJD22|away is now known as RJD22 | 04:05 | |
| *** dendrobates is now known as dendro-afk | 04:06 | |
| *** adiantum has quit IRC | 04:09 | |
| *** kazu has quit IRC | 04:16 | |
| *** adiantum has joined #openstack | 04:16 | |
| *** dirakx has joined #openstack | 04:18 | |
| *** adiantum has quit IRC | 04:23 | |
| *** adjohn has joined #openstack | 04:26 | |
| *** adiantum has joined #openstack | 04:29 | |
| *** pvo has quit IRC | 04:43 | |
| *** dendro-afk is now known as dendrobates | 04:45 | |
| *** lstoll has quit IRC | 04:53 | |
| *** lstoll has joined #openstack | 04:54 | |
| *** lamar has quit IRC | 04:57 | |
| *** lamar has joined #openstack | 05:01 | |
| *** littleidea has joined #openstack | 05:05 | |
| *** h1nchtastic has quit IRC | 05:10 | |
| *** h1nch has joined #openstack | 05:10 | |
| *** annegentle has quit IRC | 05:10 | |
| *** annegentle has joined #openstack | 05:11 | |
| *** fysa has quit IRC | 05:11 | |
| *** lamar has quit IRC | 05:24 | |
| *** adiantum has quit IRC | 05:25 | |
| *** dendrobates is now known as dendro-afk | 05:30 | |
| *** adiantum has joined #openstack | 05:30 | |
| *** localhost3 has quit IRC | 05:34 | |
| *** f4m8_ is now known as f4m8 | 05:45 | |
| *** littleidea has quit IRC | 05:49 | |
| *** fysa has joined #openstack | 05:54 | |
| *** adiantum has quit IRC | 05:57 | |
| *** MarkAtwood has joined #openstack | 05:58 | |
| *** adiantum has joined #openstack | 06:02 | |
| *** zenmatt has quit IRC | 06:17 | |
| *** littleidea has joined #openstack | 06:27 | |
| *** littleidea_ has joined #openstack | 06:46 | |
| *** littleidea has quit IRC | 06:46 | |
| *** littleidea_ is now known as littleidea | 06:46 | |
| *** MarkAtwood has quit IRC | 06:49 | |
| *** miclorb has quit IRC | 06:50 | |
| *** adiantum has quit IRC | 06:55 | |
| *** MarkAtwood has joined #openstack | 06:56 | |
| *** guigui has joined #openstack | 06:57 | |
| *** hazmat has joined #openstack | 06:58 | |
| *** adiantum has joined #openstack | 07:00 | |
| *** Manikandan has joined #openstack | 07:01 | |
| Manikandan | hi all, in openstack compute i have to install both compute and client separately | 07:03 |
| *** Nacx has joined #openstack | 07:04 | |
| *** adiantum has quit IRC | 07:08 | |
| *** brd_from_italy has quit IRC | 07:08 | |
| *** mgoldmann has joined #openstack | 07:09 | |
| *** hazmat has quit IRC | 07:10 | |
| *** adiantum has joined #openstack | 07:13 | |
| *** fysa_ has joined #openstack | 07:16 | |
| *** fysa has quit IRC | 07:18 | |
| *** fysa_ is now known as fysa | 07:18 | |
| *** naehring has joined #openstack | 07:22 | |
| *** slyphon has quit IRC | 07:25 | |
| *** reldan has joined #openstack | 07:28 | |
| *** masumotok has joined #openstack | 07:29 | |
| *** kashyap has joined #openstack | 07:30 | |
| *** mahadev has quit IRC | 07:33 | |
| *** fysa has quit IRC | 07:34 | |
| *** gasbakid has joined #openstack | 07:36 | |
| *** adiantum has quit IRC | 07:39 | |
| *** adiantum has joined #openstack | 07:40 | |
| *** fysa has joined #openstack | 07:41 | |
| *** rcc has joined #openstack | 07:43 | |
| *** allsystemsarego has joined #openstack | 07:44 | |
| *** gregp76_ has joined #openstack | 07:46 | |
| *** mdomsch has quit IRC | 07:51 | |
| *** reldan has quit IRC | 07:53 | |
| *** gregp76_ has quit IRC | 07:58 | |
| *** MarkAtwood has quit IRC | 08:09 | |
| *** noguchi has joined #openstack | 08:10 | |
| *** mahadev has joined #openstack | 08:11 | |
| *** littleidea has quit IRC | 08:21 | |
| *** 16WAAKRE3 has joined #openstack | 08:29 | |
| *** ewindisch has joined #openstack | 08:29 | |
| *** hazmat has joined #openstack | 08:30 | |
| *** 16WAAKRE3 has quit IRC | 08:30 | |
| *** ewindisch has quit IRC | 08:30 | |
| *** ewindisch has joined #openstack | 08:30 | |
| *** ewindisch has quit IRC | 08:32 | |
| *** mgoldmann has quit IRC | 08:34 | |
| *** miclorb has joined #openstack | 08:35 | |
| *** mgoldmann has joined #openstack | 08:36 | |
| *** DigitalFlux has joined #openstack | 08:37 | |
| *** ramkrsna has joined #openstack | 08:43 | |
| *** ramkrsna has joined #openstack | 08:43 | |
| *** Oneiropolo has joined #openstack | 08:45 | |
| Oneiropolo | i have a question for about node fail over | 08:47 |
| Oneiropolo | is there somebody can help? | 08:47 |
| *** mahadev has quit IRC | 08:47 | |
| *** daveiw has joined #openstack | 08:48 | |
| kpepple | Oneiropolo: what do you mean by node fail over ? | 08:48 |
| Oneiropolo | when a server is broke down | 08:49 |
| kpepple | Oneiropolo: yeeesss :) | 08:49 |
| Oneiropolo | swift is using consistent hashing ring | 08:49 |
| kpepple | Oneiropolo: ahhh ... you are talking about swift ... i thought you were talking about nova (we have a new live migration feature coming) | 08:50 |
| kpepple | Oneiropolo: probably can't help you on the swift one ... | 08:50 |
| Oneiropolo | oh, that sounds cool | 08:50 |
| Oneiropolo | nova gonna have new feature . | 08:50 |
| Oneiropolo | anyway, my question is | 08:50 |
| Oneiropolo | when a object node is fail. | 08:50 |
| Oneiropolo | there are some replications for that node data. | 08:51 |
| Oneiropolo | at that time, those replications serve the request directly? | 08:51 |
| Oneiropolo | or just give copy of the objects to eligible server? | 08:52 |
| Oneiropolo | I know that swift is using consistent hashing ring. | 08:52 |
| Oneiropolo | that means that when a node on the ring is fail, | 08:52 |
| Oneiropolo | it doesn't affect other node as much | 08:52 |
| *** slyphon has joined #openstack | 08:53 | |
| Oneiropolo | so in the theory of consistent hashing | 08:53 |
| Oneiropolo | that means the objects which was fail node had should copied to node that right location of the ring. | 08:54 |
| Oneiropolo | i wanna know how swift handle this situation. | 08:55 |
| *** mikemc_home has joined #openstack | 08:58 | |
| kpepple | Oneiropolo: AFAIK, the ring should contain a mapping of object to the 3 partition / physical locations and it will handle the failure by pointing client to another (hopefully alive) location. is this what you are asking ? | 08:58 |
| Oneiropolo | yes | 08:59 |
| Oneiropolo | so, you mean when a node is fail, different node gonna serve that request? | 08:59 |
| kpepple | Oneiropolo: yes, the ring should point the to a different node | 09:01 |
| Oneiropolo | depend on consistent hashing algorithm | 09:01 |
| kpepple | Oneiropolo: yes | 09:02 |
| Oneiropolo | the object should go to right node. | 09:02 |
| Oneiropolo | I don't know whether the replicas located in right place (depend on consistent hashing - on the ring) or just copied to right place | 09:04 |
| kpepple | Oneiropolo: i am not sure on that | 09:04 |
| Oneiropolo | okay. | 09:04 |
| Oneiropolo | what is your guess? | 09:05 |
| Oneiropolo | maybe i need to check out sources but i don't have enough time now . | 09:05 |
| kpepple | Oneiropolo: from reading this (http://swift.openstack.org/overview_architecture.html), it appears that it is configurable | 09:06 |
| *** Oneiropolo_ has joined #openstack | 09:07 | |
| Oneiropolo_ | i'm back | 09:07 |
| *** MarcMorata has joined #openstack | 09:07 | |
| *** adjohn has quit IRC | 09:08 | |
| *** magglass1 has quit IRC | 09:09 | |
| *** j05h has quit IRC | 09:10 | |
| *** j05h has joined #openstack | 09:10 | |
| *** magglass1 has joined #openstack | 09:11 | |
| *** slyphon has quit IRC | 09:13 | |
| *** arun has quit IRC | 09:15 | |
| *** arun has joined #openstack | 09:15 | |
| *** arun has joined #openstack | 09:15 | |
| *** adiantum has quit IRC | 09:24 | |
| *** arun_ has quit IRC | 09:29 | |
| *** adiantum has joined #openstack | 09:31 | |
| *** irahgel has joined #openstack | 09:36 | |
| Oneiropolo_ | i wanna ask about swift ~ | 09:37 |
| wedjat | just ask | 09:40 |
| Oneiropolo_ | thank you wedjat | 09:52 |
| Oneiropolo_ | when a server node is fail | 09:52 |
| Oneiropolo_ | like object | 09:52 |
| Oneiropolo_ | how swift handle this situation ? | 09:52 |
| Oneiropolo_ | i thought that | 09:53 |
| Oneiropolo_ | it use the consistent hashing | 09:53 |
| Oneiropolo_ | so, when a node is fail | 09:53 |
| Oneiropolo_ | objects that the node served should move from replication server to service node | 09:54 |
| *** gasbakid has quit IRC | 10:01 | |
| *** uksysadmin has joined #openstack | 10:04 | |
| Oneiropolo_ | on swift, | 10:06 |
| Oneiropolo_ | devices can added to the ring during service running? | 10:06 |
| *** daveiw has quit IRC | 10:16 | |
| *** Manikandan has quit IRC | 10:22 | |
| uksysadmin | just updated to the latest nova from ppa: 2011.2~bzr757 and now getting ProgrammingError "You have an error in your SQL syntax ... (near ) \n LIMIT 0, 1' at line 3) it looks like some quotes are missing around %s... anybody else get this? | 10:23 |
| *** adiantum has quit IRC | 10:26 | |
| *** reldan has joined #openstack | 10:26 | |
| *** miclorb has quit IRC | 10:34 | |
| uvirtbot | New bug: #728342 in nova "ProgrammingError "You have an error in your SQL syntax ... (near ) \n LIMIT 0, 1' at line 3)" [Undecided,New] https://launchpad.net/bugs/728342 | 10:36 |
| *** adiantum has joined #openstack | 10:38 | |
| *** irahgel has quit IRC | 10:50 | |
| *** noguchi has left #openstack | 10:51 | |
| ttx | uksysadmin: i'll try to reproduce that | 10:55 |
| uksysadmin | ttx: cheers | 10:58 |
| ttx | I reproduce it. | 11:02 |
| *** cascone has quit IRC | 11:02 | |
| *** dragondm has joined #openstack | 11:04 | |
| *** skiold has joined #openstack | 11:07 | |
| ttx | uksysadmin: confirmed, that's a critical regression, probably introduced at bzr757 | 11:07 |
| * ttx looks | 11:07 | |
| ttx | I lack the sqlalchemy foo to understand why it fails... the definitions seem correct | 11:12 |
| ttx | I reproduce on sqlite, fwiw | 11:12 |
| ttx | kpepple, jk0: ^^ | 11:12 |
| *** h0cin has joined #openstack | 11:12 | |
| *** adiantum has quit IRC | 11:12 | |
| *** adiantum has joined #openstack | 11:19 | |
| *** irahgel has joined #openstack | 11:21 | |
| ttx | hrm, looks like this would be wrong: | 11:23 |
| ttx | type_data = instance_types.get_instance_type([inst['instance_type']]) | 11:23 |
| ttx | should be: | 11:23 |
| ttx | type_data = instance_types.get_instance_type(inst['instance_type']) | 11:23 |
| ttx | but then I hit another issue. | 11:23 |
| *** Oneiropolo_ has quit IRC | 11:26 | |
| *** Oneiropolo has quit IRC | 11:26 | |
| ttx | more after lunch. | 11:28 |
| BK_man | guys, what version of boto do you have in Ubuntu? | 11:33 |
| *** j05h has quit IRC | 11:33 | |
| *** rds__ has joined #openstack | 11:34 | |
| *** skiold has quit IRC | 11:36 | |
| uksysadmin | ttx: noted, ta. | 11:38 |
| soren | python-boto | 1.9b-1ubuntu3 | lucid | source, all | 11:38 |
| soren | python-boto | 1.9b-1ubuntu3 | maverick | source, all | 11:38 |
| soren | python-boto | 1.9b-1ubuntu4 | natty | source, all | 11:38 |
| soren | BK_man: ^ | 11:38 |
| BK_man | soren: thanks. That's not a case for me. I really need to figure out which version of euca2ools I need to install to use it with Nova | 11:39 |
| soren | euca2ools | 1.2-0ubuntu10 | lucid | source, all | 11:39 |
| soren | euca2ools | 1.2-0ubuntu10.1 | lucid-updates | source, all | 11:39 |
| soren | euca2ools | 1.2-0ubuntu11 | maverick | source, all | 11:39 |
| soren | euca2ools | 1.2-0ubuntu11.1 | maverick-updates | source, all | 11:39 |
| soren | euca2ools | 1.3.1-0ubuntu6 | natty | source, all | 11:39 |
| soren | Those are known to work. | 11:39 |
| BK_man | soren: 1.3.1 fails with euca-get-ajax-console claiming that there is no EC2 attribute 'get-ajax-console', 1.2 if failing due non-existent class ConnectionFailed in euca2ools/__init__.py | 11:40 |
| soren | Ah. | 11:40 |
| BK_man | soren: could you please simply run euca-get-ajax-proxy and tell me your output and version of euca2ools? | 11:41 |
| BK_man | problematic string in euca-get-ajax-console is: from euca2ools import Euca2ool, InstanceValidationError, Util, ConnectionFailed | 11:41 |
| *** guigui has quit IRC | 11:42 | |
| *** allsystemsarego has quit IRC | 11:43 | |
| *** adiantum has quit IRC | 11:46 | |
| uksysadmin | ttx: nova/virt/libvirt_conn.py already has that change you noted : type_data = instance_types.get_instance_type([inst['instance_type']]) | 11:46 |
| soren | BK_man: tools/euca-get-ajax-console from current nova trunk gives me: http://paste.ubuntu.com/574935/ | 11:48 |
| BK_man | soren: this is using which version of euca2ools? | 11:52 |
| soren | 1.3.1-0ubuntu6 | 11:53 |
| BK_man | soren: ok. could you please try to specify instance id? | 11:54 |
| *** j05h has joined #openstack | 11:55 | |
| BK_man | soren: this is what i got: | 11:55 |
| BK_man | EC2Connection instance has no attribute 'get_ajax_console' | 11:55 |
| soren | I do too. | 11:55 |
| soren | Ah. | 11:58 |
| *** sateesh has joined #openstack | 11:59 | |
| sateesh | Can some one point me to nova build for 2011.1.1 | 11:59 |
| BK_man | sateesh: it wan't released yet. You probably want to download a snapshot here: http://nova.openstack.org/tarballs/?C=M;O=D | 12:00 |
| sateesh | BK_man: thanks, do you when is it scheduled for release? | 12:03 |
| BK_man | sateesh: it should be released today. ask ttx for details | 12:03 |
| sateesh | BK_man: thanks | 12:04 |
| uvirtbot | New bug: #728392 in nova "tools/euca-get-ajax-console fails with euca2ools 1.2 or 1.3.1" [Undecided,New] https://launchpad.net/bugs/728392 | 12:06 |
| sateesh | Anybody tried spawning multiple instances parallelly? | 12:07 |
| uksysadmin | sateesh, with -n parameter? | 12:10 |
| *** perestrelka has quit IRC | 12:11 | |
| sateesh | uksysadmin: no, running euca-run-instance <instance_id> commands simultaneously. | 12:11 |
| *** perestrelka has joined #openstack | 12:11 | |
| uksysadmin | why would you do that? | 12:11 |
| *** allsystemsarego has joined #openstack | 12:13 | |
| *** allsystemsarego has joined #openstack | 12:13 | |
| uvirtbot | New bug: #728398 in nova "tools/euca-get-ajax-console traps early without meaningful error message" [Undecided,New] https://launchpad.net/bugs/728398 | 12:16 |
| *** ctennis_ has quit IRC | 12:17 | |
| *** romain_lenglet_ has joined #openstack | 12:20 | |
| *** Nacx has quit IRC | 12:27 | |
| *** mikemc_home has quit IRC | 12:28 | |
| ttx | uksysadmin: the extra pair of braces loses sqlalchemy. | 12:33 |
| *** ctennis_ has joined #openstack | 12:34 | |
| *** sateesh has quit IRC | 12:34 | |
| uksysadmin | ah - I see | 12:35 |
| uksysadmin | I've just changed this and now get another error | 12:35 |
| uksysadmin | which looks like a log file permissions error so gonna see why that's the case. at least I don't get the SQL error now. | 12:37 |
| ttx | uksysadmin: "chardev: opening backend "file" failed" ? | 12:37 |
| uksysadmin | yeah that's the one | 12:37 |
| ttx | uksysadmin: what was the last version you successfully used ? | 12:39 |
| * ttx wonders if that's a regression in libvirt 0.8.8-1ubuntu1~ppamaverick1 | 12:40 | |
| uksysadmin | 2011.2~bzr749 | 12:40 |
| *** konquertech has joined #openstack | 12:41 | |
| *** konquertech has quit IRC | 12:42 | |
| ttx | uksysadmin: the other issue is a regression in libvirt in PPA | 12:46 |
| ttx | soren: ^ | 12:46 |
| ttx | I downgraded to 0.8.3-1ubuntu14 and it works with my patch | 12:47 |
| ttx | uksysadmin: try sudo apt-get install libvirt-bin=0.8.3-1ubuntu14 libvirt0=0.8.3-1ubuntu14 python-libvirt=0.8.3-1ubuntu14 | 12:47 |
| * ttx pushes branch to fix | 12:48 | |
| soren | ttx: Sorry, what's the issue? | 12:50 |
| ttx | uksysadmin: any chance you could file your log where soren can see it ? | 12:51 |
| ttx | soren: libvirt in maverick PPA fails to start instances | 12:51 |
| ttx | "chardev: opening backend "file" failed" | 12:51 |
| ttx | uksysadmin can probably post the full log | 12:51 |
| *** westmaas1 has quit IRC | 12:52 | |
| uksysadmin | yeah I'll get a log now... pastebin it? | 12:54 |
| ttx | yep | 12:54 |
| *** zul has joined #openstack | 12:55 | |
| *** guigui1 has joined #openstack | 12:56 | |
| *** f4m8 is now known as f4m8_ | 12:56 | |
| uksysadmin | soren, ttx: http://paste.openstack.org/show/792/ | 12:58 |
| uksysadmin | I'll downgrade as suggested | 12:58 |
| soren | I'll be able to reproduce in a minute. | 12:59 |
| ttx | uksysadmin: that's what I got too, thanks for the paste | 12:59 |
| uksysadmin | np | 12:59 |
| ttx | soren: you might need my fix to bug 728342 to reach that point. | 12:59 |
| uvirtbot | Launchpad bug 728342 in nova "Regression prevents to start any instance through libvirt" [Critical,In progress] https://launchpad.net/bugs/728342 | 12:59 |
| soren | ttx: Awesomesauce. | 13:00 |
| *** ramkrsna has quit IRC | 13:01 | |
| ttx | nova-core: regression fix proposed as https://code.launchpad.net/~ttx/nova/lp728342/+merge/52061 -- give it some love before it blocks someone else | 13:02 |
| uksysadmin | yeah confirmed the downgrade now works (with the extra square brackets removed in libvirt_conn.py for type_data = instance_types.get_instance_type... | 13:03 |
| * ttx gets back to his regular schedule: releasing 2011.1.1 | 13:03 | |
| BK_man | ttx: release it, please :) | 13:03 |
| ttx | BK_man: before you find another bug :) | 13:04 |
| * uksysadmin will also refrain from an apt-get update. Every little helps an' all that ;-) | 13:04 | |
| *** BK_man has quit IRC | 13:07 | |
| *** BK_man has joined #openstack | 13:07 | |
| *** kashyap has quit IRC | 13:07 | |
| *** omidhdl has joined #openstack | 13:11 | |
| uksysadmin | would be nice to try that new version of libvirt as every now and again I get: libvirtError: internal error Process exited while reading console log output: char device redirected to /dev/pts/3 | 13:12 |
| uksysadmin | seems suspiciously similar to that newer version's error | 13:13 |
| *** ctennis_ has quit IRC | 13:14 | |
| *** ctennis_ has joined #openstack | 13:14 | |
| *** ctennis has left #openstack | 13:18 | |
| ttx | ok, it's out. | 13:19 |
| *** annegentle has quit IRC | 13:20 | |
| BK_man | ttx: thanks! | 13:21 |
| * uksysadmin starts a polite applause expecting others to follow | 13:21 | |
| *** annegentle has joined #openstack | 13:21 | |
| * iRTermite applauses like the sheep he's been requested to be. ;) | 13:22 | |
| uksysadmin | that was one extra I was expecting | 13:24 |
| DigitalFlux | What's up with sudo on the compute workers ?! | 13:27 |
| DigitalFlux | (nova.root): TRACE: Stderr: 'sudo: no tty present and no askpass program specified\n' | 13:27 |
| ttx | DigitalFlux: usually this points to a broken nova_sudoers file | 13:27 |
| DigitalFlux | uksysadmin: Hi man | 13:27 |
| ttx | DigitalFlux: what user are you running nova under ? nova ? root ? | 13:28 |
| DigitalFlux | ttx: Could you please give me a sudoers file that works .. | 13:28 |
| DigitalFlux | it's root i guess | 13:28 |
| DigitalFlux | checking .. | 13:28 |
| *** kazu has joined #openstack | 13:28 | |
| soren | DigitalFlux: What does the preceding lines say? | 13:28 |
| DigitalFlux | oh first time to notice that | 13:28 |
| uksysadmin | DigitalFlux, g'day. | 13:28 |
| DigitalFlux | soren: it's the 'nova' user | 13:28 |
| DigitalFlux | soren: it's a long dnsmasq command | 13:29 |
| DigitalFlux | soren: ttx: how can i just go and run it as root ? | 13:29 |
| DigitalFlux | config available for that ? | 13:29 |
| *** MarcMorata has quit IRC | 13:29 | |
| soren | ttx: Darn it, this is a libvirt bug. | 13:29 |
| DigitalFlux | uksysadmin: good day man :) | 13:29 |
| DigitalFlux | soren: :(, another libvirt bug ? | 13:30 |
| ttx | DigitalFlux: I think that's a bad idea, better fix your nova_sudoers setup | 13:30 |
| *** ironcamel has quit IRC | 13:30 | |
| DigitalFlux | ttx: and where is that ? :) | 13:30 |
| ttx | DigitalFlux: /etc/sudoers.d/nova_sudoers | 13:30 |
| DigitalFlux | ttx: checking .. | 13:30 |
| ttx | DigitalFlux: (in Ubuntu packaging) | 13:30 |
| BK_man | DigitalFlux: https://github.com/abrindeyev/openstack-nova-rhel6/blob/bexar/SOURCES/openstack-nova-sudoers | 13:31 |
| *** ironcamel has joined #openstack | 13:31 | |
| DigitalFlux | ttx: BK_man: I think this is identical to mine | 13:31 |
| sandywalsh | "The diff has been truncated for viewing." ... wuh?! damn you launchpad. | 13:31 |
| sandywalsh | :p | 13:31 |
| ttx | sandywalsh: it's for your owngood :) | 13:32 |
| DigitalFlux | so i should just enter that dnsmasq command here i guess | 13:32 |
| soren | DigitalFlux: No, same one. | 13:32 |
| BK_man | DigitalFlux: mine should be different - it's from RHEL6 and carefully checked line by libe | 13:32 |
| BK_man | s/libe/line/ | 13:32 |
| DigitalFlux | it's already there :( | 13:32 |
| ttx | DigitalFlux: does your /etc/sudoers #include /etc/sudoers.d ? | 13:32 |
| DigitalFlux | ttx: OH ! :( | 13:33 |
| DigitalFlux | ttx: The default suderos file was replaced by puppet | 13:33 |
| uksysadmin | as we seem to be picking on libvirt today - that error I posted earlier regarding reading console log output: char device redirected to /dev/pts/3 seems to be related out of resource (ram)... basically fails ungracefully. | 13:33 |
| BK_man | DigitalFlux: echo '#includedir /etc/sudoers.d' >> /etc/sudoers | 13:33 |
| DigitalFlux | ttx: and the include statement is hashed already | 13:33 |
| ttx | DigitalFlux: bad puppet. | 13:33 |
| DigitalFlux | BK_man: I think i should uncomment it, no ? | 13:33 |
| BK_man | DigitalFlux: no | 13:34 |
| DigitalFlux | BK_man: then it's already there .. | 13:34 |
| *** MarcMorata has joined #openstack | 13:34 | |
| BK_man | DigitalFlux: try it - give user nova shell, su - nova and execute sudo -l | 13:34 |
| soren | ?!? | 13:34 |
| soren | 14:33:49.111: 18995: info : virSecurityDACSetOwnership:99 : Setting DAC user and group on '/var/lib/nova/instances/instance-00000004/console.log' to '18995:1' | 13:34 |
| *** adiantum has joined #openstack | 13:37 | |
| *** iammartian has quit IRC | 13:38 | |
| DigitalFlux | BK_man: yeah, i see that sudo -l lists that nova should do dnsmasq | 13:38 |
| DigitalFlux | however i still have the same issue | 13:39 |
| DigitalFlux | not solved | 13:39 |
| *** arefaey has joined #openstack | 13:39 | |
| BK_man | DigitalFlux: which issue? | 13:39 |
| DigitalFlux | the 'Stderr: 'sudo: no tty present and no askpass program specified\n'' thingy | 13:39 |
| sandywalsh | what's the correct way to init a logger in a module now? I thought it had to be more specific than nova.log.basicConfig()? | 13:41 |
| BK_man | DigitalFlux: ah, ok | 13:41 |
| BK_man | DigitalFlux: so where is your include in /etc/sudoers? | 13:41 |
| BK_man | DigitalFlux: mine is at the end | 13:41 |
| zul | i would appreciate if someone reviews nova-lxc please :) | 13:41 |
| BK_man | DigitalFlux: so in default config there is requiretty in the top of /etc/sudoers file and following include at the end of that file disables it specifically for user nova | 13:42 |
| DigitalFlux | BK_man: i don't have this requiretty line | 13:45 |
| DigitalFlux | BK_man: where should it go and what is the line exactly ? | 13:45 |
| BK_man | DigitalFlux: $ sudo grep requiretty /etc/sudoers | 13:46 |
| BK_man | Defaults requiretty | 13:46 |
| BK_man | DigitalFlux: $ sudo grep requiretty /etc/sudoers.d/openstack-nova | 13:46 |
| BK_man | Defaults:nova !requiretty | 13:46 |
| DigitalFlux | BK_man: ok just to confirm as i forgot the sudoers file format :), it's #include or include ? | 13:46 |
| *** matclayton has joined #openstack | 13:47 | |
| BK_man | DigitalFlux: sudo grep include /etc/sudoers | 13:47 |
| BK_man | #includedir /etc/sudoers.d | 13:47 |
| *** mgoldmann has quit IRC | 13:50 | |
| *** Nacx has joined #openstack | 13:51 | |
| *** mgoldmann has joined #openstack | 13:51 | |
| *** irahgel has quit IRC | 13:53 | |
| *** ramkrsna has joined #openstack | 13:53 | |
| *** ramkrsna has joined #openstack | 13:53 | |
| *** skiold has joined #openstack | 14:04 | |
| *** Ephur has joined #openstack | 14:06 | |
| DigitalFlux | I still can't figure it out | 14:10 |
| DigitalFlux | the files sames sane to me | 14:10 |
| DigitalFlux | pastebinning them .. | 14:10 |
| *** omidhdl has left #openstack | 14:10 | |
| DigitalFlux | Here is the sudoers file -> http://pastebin.com/zNpYL00X cc: BK_man, soren, ttx | 14:11 |
| *** stewart has quit IRC | 14:11 | |
| *** stewart has joined #openstack | 14:11 | |
| DigitalFlux | And here is nova_sudoers -> http://pastebin.com/5YBNfB3D | 14:12 |
| *** mray has joined #openstack | 14:13 | |
| ttx | DigitalFlux: sounds alright to me -- maybe try to restart a few things to get the /etc/sudoers change taken into account | 14:13 |
| DigitalFlux | ttx: doing the restart loop .. | 14:14 |
| creiht | doh, Oneiropolo is gone already :/ | 14:15 |
| *** dprince has joined #openstack | 14:17 | |
| DigitalFlux | ttx: same error :( | 14:17 |
| *** gasbakid has joined #openstack | 14:19 | |
| DigitalFlux | ttx: I even replaced the sudoers file with the default that installs with Ubuntu server | 14:20 |
| *** vvuksan has joined #openstack | 14:20 | |
| *** gondoi has joined #openstack | 14:21 | |
| *** irahgel has joined #openstack | 14:21 | |
| DigitalFlux | ttx: can u guide me on how to run nova as root for the moment | 14:21 |
| DigitalFlux | ttx: just for a testing installation | 14:21 |
| ttx | DigitalFlux: never tried, but I guess it's just a matter of changing the user starting nova. | 14:22 |
| DigitalFlux | ttx: hmm, checking the init scripts | 14:22 |
| DigitalFlux | so the commands is nova-volume --flagfile=/etc/nova/nova.conf nova | 14:27 |
| DigitalFlux | that 2nd argument is the username ? | 14:27 |
| *** rchavik has quit IRC | 14:31 | |
| DigitalFlux | OK guys i surrender :) soren: ttx: BK_man | 14:31 |
| DigitalFlux | Nova-compute and volume is operating using the root user | 14:31 |
| DigitalFlux | and i still get those ssh-askpass errors | 14:31 |
| *** zenmatt has joined #openstack | 14:35 | |
| soren | libvirt bug found and fixed. | 14:36 |
| soren | Will upload in a few minutes if my tests work out. | 14:36 |
| DigitalFlux | soren: this bug is related to my problem ? | 14:36 |
| soren | DigitalFlux: Your sudo problem? No. | 14:37 |
| DigitalFlux | soren: ok | 14:37 |
| soren | DigitalFlux: Are you sure they | 14:37 |
| soren | re running as root? | 14:37 |
| DigitalFlux | soren: here is the o/p of the process list | 14:38 |
| DigitalFlux | soren: http://pastebin.com/FCEC73Hc | 14:38 |
| soren | DigitalFlux: And can you paste a full traceback of the error, please? | 14:38 |
| DigitalFlux | soren: ok ... | 14:38 |
| *** ppetraki has joined #openstack | 14:39 | |
| DigitalFlux | soren: here it is -> http://pastebin.com/w2gViL35 | 14:39 |
| soren | DigitalFlux: Multi-node install? | 14:42 |
| DigitalFlux | soren: yup | 14:42 |
| DigitalFlux | soren: just 2 nodes | 14:42 |
| soren | DigitalFlux: it's a nova-network problem. | 14:42 |
| soren | Wrong box. | 14:42 |
| soren | it just passes the error back through rpc. | 14:42 |
| DigitalFlux | DigitalFlux: hmm | 14:43 |
| *** guynaor has joined #openstack | 14:46 | |
| *** guynaor has left #openstack | 14:47 | |
| ttx | soren: cool | 14:50 |
| DigitalFlux | Unexpected error while running command. Command: sudo brctl addbr br100 | 14:50 |
| *** irahgel has quit IRC | 14:51 | |
| *** zul has quit IRC | 14:54 | |
| *** mdomsch has joined #openstack | 14:56 | |
| *** irahgel has joined #openstack | 14:56 | |
| DigitalFlux | nova-network is indeed having problems with ssh-askpass | 14:58 |
| DigitalFlux | ok i will make nova on the controller node too start as root | 14:59 |
| openstackhudson | Project swift build #209: SUCCESS in 37 sec: http://hudson.openstack.org/job/swift/209/ | 15:01 |
| openstackhudson | Tarmac: s3api: use boto to get canonical string for signature | 15:01 |
| openstackhudson | Replace the homegrown function to get a canonical string for signature. | 15:01 |
| *** zul has joined #openstack | 15:03 | |
| soren | ttx: libvirt fix accepted into natty proper, and the maverick and lucid ppa's. | 15:03 |
| *** littleidea has joined #openstack | 15:04 | |
| ttx | soren: cool | 15:05 |
| ttx | trying that now | 15:05 |
| ttx | hm, I'll let it build first | 15:06 |
| *** hazmat has quit IRC | 15:08 | |
| *** mustfg has joined #openstack | 15:09 | |
| soren | natty will be a while before it's published. maverick and lucid not entirely as much. | 15:09 |
| *** skiold has quit IRC | 15:10 | |
| *** reldan has quit IRC | 15:12 | |
| *** skiold has joined #openstack | 15:15 | |
| mustfg | i can someone help me with ssh ´ing into instances created in VLAN mode ? how can i do that ? | 15:17 |
| vishy | mustfg: what is the problem? | 15:19 |
| *** dendro-afk is now known as dendrobates | 15:21 | |
| *** reldan has joined #openstack | 15:22 | |
| *** hazmat has joined #openstack | 15:23 | |
| ttx | vishy: please have a look at https://code.launchpad.net/~ttx/nova/lp728342/+merge/52061 -- should be ready to go if you're OK with it | 15:24 |
| *** mahadev has joined #openstack | 15:25 | |
| soren | ttx: It already has two +1's. | 15:25 |
| vishy | ttx: saw it, didn't mention anything because there are already two approves | 15:25 |
| soren | ttx: I've just approved it. | 15:25 |
| ttx | soren: ack, thanks | 15:25 |
| *** blueadept has joined #openstack | 15:26 | |
| vishy | soren: so you found another bug in libvirt? | 15:27 |
| *** MarkAtwood has joined #openstack | 15:27 | |
| *** msassak has joined #openstack | 15:27 | |
| soren | vishy: Found, fixed, uploaded, sent upstream. | 15:28 |
| *** MarkAtwood has quit IRC | 15:28 | |
| soren | ..and accepted upstream, apparantly. | 15:28 |
| *** MarkAtwood has joined #openstack | 15:28 | |
| soren | http://libvirt.org/git/?p=libvirt.git;a=commit;h=e5f3b90e975a21222985641abb4ebaef0e616714 if anyone should care. | 15:29 |
| vishy | soren: cool. I still haven't gotten lxc to work properly yet. But I haven't had time to give it a real try since the new libvirt showed up in the ppa | 15:29 |
| ttx | soren: fix worksforme | 15:29 |
| soren | \o/ | 15:30 |
| *** mahadev has quit IRC | 15:31 | |
| *** littleidea has quit IRC | 15:31 | |
| *** jero has joined #openstack | 15:31 | |
| kpepple | ttx: just getting up ... and seeing the regression. what needs to be done ? | 15:31 |
| ttx | kpepple: should be fixed now | 15:31 |
| ttx | a few more minutes while hudson digests it | 15:32 |
| kpepple | ttx: is that lp728342 ? | 15:32 |
| ttx | yes | 15:32 |
| *** slyphon has joined #openstack | 15:32 | |
| *** westmaas has quit IRC | 15:32 | |
| soren | It's in. | 15:32 |
| *** grapex has joined #openstack | 15:33 | |
| ttx | kpepple: nothing like getting your regressions fixed while you sleep :) | 15:33 |
| *** mahadev has joined #openstack | 15:33 | |
| *** hvaldivia has joined #openstack | 15:34 | |
| kpepple | ttx: it is my preferred method ... but not the best way to wake up :( | 15:34 |
| hvaldivia | Hi everybody | 15:34 |
| hvaldivia | I have a question... | 15:34 |
| *** Vek has quit IRC | 15:35 | |
| ttx | hvaldivia: that's good, because we sometimes have answers. | 15:35 |
| hvaldivia | hahaha. wait I'm thinking how to ask my question | 15:36 |
| kpepple | ttx: just as a bit of postmortem ... i see why this wasn't caught in unittest (due to the fakes) but why didn't smoketests catch this ? i thought we launched instances there | 15:36 |
| soren | kpepple: We do, but they're not run prior to merge. | 15:37 |
| hvaldivia | If I set --fixed_range=10.0.0.0/12 | 15:37 |
| hvaldivia | it means that my vms will have ips in that range (10.0.0.0/12) | 15:37 |
| soren | kpepple: Yet, at least. | 15:37 |
| openstackhudson | Project nova build #596: SUCCESS in 1 min 46 sec: http://hudson.openstack.org/job/nova/596/ | 15:37 |
| openstackhudson | Tarmac: Fix regression in the way libvirt_conn gets its instance_types | 15:37 |
| hvaldivia | later What does it mean: nova-manage network create 192.168.0.0/24 1 255 | 15:38 |
| hvaldivia | Can I use nova-manage network create 10.0.0.0/24 1 255 | 15:38 |
| hvaldivia | What is the differences? | 15:38 |
| vishy | hvaldivia: no | 15:38 |
| kpepple | soren: where can we access the smoketest run results -- is that the same hudson/jenkins site ? | 15:38 |
| vishy | 10.0.0.0/24 1 256 will work though | 15:39 |
| soren | kpepple: Not yet. | 15:39 |
| vishy | kpepple: smoketests aren't run automatically on trunk yet | 15:39 |
| kpepple | soren , vishy: so as a followup ... how do i execute smoketests locally ? | 15:39 |
| soren | kpepple: I used to run some smokte tests on trunk, but trunk needs to be fixed before they can start to run again. | 15:39 |
| hvaldivia | vishy: ok but hat does it mean: nova-manage network create 192.168.0.0/24 1 255 | 15:39 |
| hvaldivia | ? | 15:39 |
| *** hazmat has quit IRC | 15:39 | |
| vishy | kpepple: you need to run the system and source an rc file | 15:40 |
| *** mustfg has quit IRC | 15:40 | |
| hvaldivia | why no use 10.0.0.0/24 1 256. Here the tutorial http://docs.openstack.org/openstack-compute/admin/content/ch03s03.html | 15:40 |
| hvaldivia | I am a little confused | 15:40 |
| vishy | kpepple: highly recommend these smoketests: lp:~anso/nova/smoketests_fixes | 15:40 |
| vishy | with those you can cd <checkout>/smoketests/ | 15:41 |
| vishy | and do python run_tests.py | 15:41 |
| kpepple | vishy: do i run those on a driver systems against the running test system ? | 15:42 |
| vishy | hvaldivia: nework create should always use a subset of fixed_range if you want to avoid issues | 15:42 |
| vishy | kpepple: well you can run it locally from the network host | 15:42 |
| kpepple | vishy: ok | 15:43 |
| vishy | kpepple: if you want to run them remotely, you need to add a route to get in to the instances private ips and security group rules | 15:43 |
| *** slyphon has quit IRC | 15:43 | |
| *** mahadev has quit IRC | 15:43 | |
| vishy | e.g. route add -net <private_range> gw <network host ip> | 15:44 |
| hvaldivia | vishy: ok I think, I understood. | 15:44 |
| vishy | euca-authorize -P tcp -p 22 default | 15:44 |
| *** mdomsch has quit IRC | 15:44 | |
| vishy | euca-authorize -P icmp -t -1:-1 default | 15:44 |
| *** mustfg has joined #openstack | 15:46 | |
| *** mray has quit IRC | 15:50 | |
| *** mahadev has joined #openstack | 15:50 | |
| rds__ | Hi guys, I'm having some problem with nova-compute(after libvirt upgrade from ppa), here's the error http://pastebin.com/9jQzqvBx | 15:53 |
| *** imsplitbit has joined #openstack | 15:53 | |
| rds__ | anyone who can help? | 15:53 |
| rds__ | thanks | 15:53 |
| vishy | rds__: yes libvirt bug, soren just patched it | 15:56 |
| *** mray has joined #openstack | 15:56 | |
| rds__ | vishy, ok, thanks :) | 15:57 |
| *** mdomsch has joined #openstack | 15:57 | |
| soren | rds__: It's in the ppa already. | 15:59 |
| *** naehring has quit IRC | 16:00 | |
| kpepple | vishy: anything i need to source before i run your smoketests | 16:00 |
| * soren wanders off for a while | 16:00 | |
| *** dendrobates is now known as dendro-afk | 16:02 | |
| vishy | kpepple: yes, a normal rc file | 16:03 |
| *** maplebed has quit IRC | 16:04 | |
| *** reldan has quit IRC | 16:05 | |
| *** ccustine has joined #openstack | 16:06 | |
| rds__ | soren, I'm upgrading it now, thanks | 16:06 |
| jk0 | ttx: got a bug link? | 16:07 |
| jk0 | I'll look at it now | 16:07 |
| ttx | jk0: the fix is in -- was bug 728342 | 16:08 |
| jk0 | oh good | 16:08 |
| jk0 | thanks | 16:08 |
| uvirtbot | ttx: Error: Could not parse data returned by Launchpad: The read operation timed out | 16:08 |
| ttx | uvirtbot: no, not this one | 16:08 |
| uvirtbot | ttx: Error: "no," is not a valid command. | 16:08 |
| ttx | :P | 16:08 |
| kpepple | vishy: i think i see my problem ... what is the user dependency -- admin user named admin ? | 16:10 |
| *** mustfg has quit IRC | 16:10 | |
| kpepple | vishy: ignore, i got it | 16:10 |
| vishy | kpepple: cool. Our integration testing for anso package builds uses those tests along with vagrant http://ansolabs.no-ip.org:9000/job/vagrant_smoketests/ | 16:12 |
| *** hazmat has joined #openstack | 16:12 | |
| *** Ryan_Lane has joined #openstack | 16:13 | |
| *** rnirmal has joined #openstack | 16:15 | |
| *** reldan has joined #openstack | 16:18 | |
| *** drico has quit IRC | 16:22 | |
| *** rds__ has quit IRC | 16:23 | |
| *** drico has joined #openstack | 16:23 | |
| vishy | what no python-argparse in the glance ppa? | 16:26 |
| *** hazmat has quit IRC | 16:28 | |
| *** kashyap has joined #openstack | 16:29 | |
| *** kazu has quit IRC | 16:29 | |
| *** arefaey has quit IRC | 16:35 | |
| *** tripson_ has quit IRC | 16:37 | |
| *** guigui1 has quit IRC | 16:37 | |
| *** dendro-afk is now known as dendrobates | 16:38 | |
| *** nelson has left #openstack | 16:39 | |
| *** jdarcy has joined #openstack | 16:40 | |
| BK_man | sleepsonthefloor: around? | 16:40 |
| *** nelson has joined #openstack | 16:42 | |
| *** hvaldivia has quit IRC | 16:43 | |
| *** perestrelka has quit IRC | 16:46 | |
| *** burris has quit IRC | 16:48 | |
| *** reldan has quit IRC | 16:49 | |
| *** gasbakid has quit IRC | 16:49 | |
| *** hvaldivia has joined #openstack | 16:49 | |
| *** uksysadmin has quit IRC | 16:49 | |
| *** gasbakid has joined #openstack | 16:50 | |
| *** neckwarmer has left #openstack | 16:54 | |
| *** pothos_ has joined #openstack | 16:56 | |
| *** imsplitbit has quit IRC | 16:57 | |
| *** viirya has quit IRC | 16:57 | |
| *** johnpur has joined #openstack | 16:57 | |
| *** ChanServ sets mode: +v johnpur | 16:57 | |
| *** viirya has joined #openstack | 16:57 | |
| *** blueadept has quit IRC | 16:57 | |
| *** pothos has quit IRC | 16:57 | |
| *** pothos_ is now known as pothos | 16:58 | |
| *** pothos has quit IRC | 17:01 | |
| *** pothos has joined #openstack | 17:01 | |
| *** zenmatt has quit IRC | 17:01 | |
| *** jaypipes has joined #openstack | 17:01 | |
| *** zenmatt has joined #openstack | 17:02 | |
| *** gdusbabek has quit IRC | 17:02 | |
| *** pandemicsyn has quit IRC | 17:02 | |
| *** gdusbabek has joined #openstack | 17:02 | |
| *** pandemicsyn has joined #openstack | 17:03 | |
| *** thatsdone has joined #openstack | 17:03 | |
| *** odyi has quit IRC | 17:07 | |
| *** odyi has joined #openstack | 17:07 | |
| *** odyi has joined #openstack | 17:07 | |
| *** nelson has quit IRC | 17:08 | |
| *** nelson has joined #openstack | 17:08 | |
| *** thatsdone has quit IRC | 17:10 | |
| openstackhudson | Project nova build #597: SUCCESS in 1 min 45 sec: http://hudson.openstack.org/job/nova/597/ | 17:13 |
| openstackhudson | Tarmac: This fix changes a tag contained in the DescribeKeyPairs response from <keypairsSet> to <keySet> so that Amazon EC2 access libraries which does more strict syntax checking can work with Nova. | 17:13 |
| *** slyphon has joined #openstack | 17:14 | |
| *** skiold has quit IRC | 17:15 | |
| slyphon | does anyone provide commercial support contracts available for swift? | 17:16 |
| slyphon | s/available// | 17:16 |
| DigitalFlux | Hi | 17:16 |
| *** spectorclan has joined #openstack | 17:17 | |
| DigitalFlux | cannot run lease-init script \\/usr\\/lib\\/pymodules\\/python2.6\\/bin\\/nova-dhcpbridge: No such file or directory\\n'\n"] | 17:20 |
| DigitalFlux | I had to install dnsmasq manually, it was not included as a dependency of nova-network | 17:20 |
| DigitalFlux | is this normal ? | 17:20 |
| *** westmaas has joined #openstack | 17:20 | |
| jarrod | digitalflux i did the same thing | 17:21 |
| DigitalFlux | jarrod: and it throws the above error ? | 17:21 |
| jarrod | oh, i received no error | 17:22 |
| DigitalFlux | hmm | 17:23 |
| jarrod | but it seems like its looking for that script in that path | 17:23 |
| jarrod | figure out why its not there | 17:23 |
| DigitalFlux | nova-dhcpbridge is located at /usr/bin/nova-dhcpbridge | 17:23 |
| DigitalFlux | not in /usr/lib/pymodules/python2.6/bin/nova-dhcpbridge | 17:23 |
| *** littleidea has joined #openstack | 17:23 | |
| DigitalFlux | Why it's even looking at this dir ! | 17:23 |
| *** joearnold has joined #openstack | 17:26 | |
| *** gondoi_ has joined #openstack | 17:27 | |
| *** dsockwell has quit IRC | 17:28 | |
| *** gondoi has quit IRC | 17:28 | |
| *** dsockwell has joined #openstack | 17:28 | |
| BK_man | DigitalFlux: --dhcpbridge=/usr/bin/nova-dhcpbridge | 17:29 |
| BK_man | DigitalFlux: put that line in /etc/nova/nova.conf | 17:29 |
| DigitalFlux | BK_man: ok thanks | 17:29 |
| DigitalFlux | BK_man: i just created a symlink actually | 17:30 |
| BK_man | DigitalFlux: see other possible flags: https://github.com/abrindeyev/openstack-nova-rhel6/blob/master/SOURCES/openstack-nova-cc-config.conf | 17:30 |
| *** Vek has joined #openstack | 17:31 | |
| *** DigitalFlux has quit IRC | 17:31 | |
| *** littleidea has quit IRC | 17:35 | |
| *** h1nchtastic has joined #openstack | 17:36 | |
| *** h1nch has quit IRC | 17:37 | |
| *** kashyap has quit IRC | 17:37 | |
| *** kashyap has joined #openstack | 17:37 | |
| *** blueadept has joined #openstack | 17:38 | |
| *** piken_ has joined #openstack | 17:38 | |
| *** joearnold has quit IRC | 17:39 | |
| *** Ryan_Lane is now known as Ryan_Lane|away | 17:40 | |
| *** sirp_ has quit IRC | 17:40 | |
| *** sirp_ has joined #openstack | 17:40 | |
| *** piken has quit IRC | 17:41 | |
| *** kainam has joined #openstack | 17:47 | |
| *** Nacx has quit IRC | 17:47 | |
| *** m_3 has quit IRC | 17:49 | |
| *** MarkAtwood has quit IRC | 17:49 | |
| *** arun has quit IRC | 17:49 | |
| *** arcane has quit IRC | 17:49 | |
| *** m_3 has joined #openstack | 17:50 | |
| *** MarkAtwood has joined #openstack | 17:51 | |
| *** arun has joined #openstack | 17:51 | |
| *** m_3 has quit IRC | 17:51 | |
| *** kashyap has quit IRC | 17:51 | |
| *** piken__ has joined #openstack | 17:51 | |
| uvirtbot | New bug: #728587 in nova "OS API /limits resource" [Undecided,In progress] https://launchpad.net/bugs/728587 | 17:51 |
| *** gregp76 has joined #openstack | 17:52 | |
| jaypipes | sirp_: I'm now done with jury duty, fyi | 17:52 |
| eday | jaypipes: guilty? :) | 17:52 |
| *** m_3 has joined #openstack | 17:52 | |
| jaypipes | eday: not guilty on all 5 counts. | 17:53 |
| *** piken_ has quit IRC | 17:54 | |
| eday | heh | 17:54 |
| *** zul has quit IRC | 17:55 | |
| jaypipes | eday: has anything been decided on Auth standardization in the past three days? | 17:56 |
| jaypipes | eday: I have re-read your email describing Swift and Nova entities and I'm still confused. | 17:57 |
| jaypipes | eday: for instance, in your first sentence after the "Nova" heading, you say "Nova currently has users and groups." | 17:57 |
| *** MarkAtwood has quit IRC | 17:57 | |
| *** arun has quit IRC | 17:57 | |
| jaypipes | eday: bu then you don't explain what a "group" is in the entire paragraph. | 17:57 |
| jaypipes | eday: in fact the term "group" isn't used at all after that first sentence. :( | 17:58 |
| *** aliguori has joined #openstack | 17:59 | |
| *** m_3 has quit IRC | 18:01 | |
| jaypipes | creiht: where can I read some developer-specific docs about auth model in Swift? I can't seem to figure it out. Some Swift URIs seem to use: http://user:key@authurl/container/objname and yet others seem to say that the URI structure is http://user:key@authurl/account/container/objname. Yet others have referred to the URI as http://user:account:key@authurl/container/objname. Which is it? | 18:01 |
| *** aryan_ has joined #openstack | 18:02 | |
| jaypipes | creiht: the above question is re: 2 bugs in Glance I'm trying to get fixed (bug #713126 and bug #717431) | 18:03 |
| uvirtbot | Launchpad bug 713126 in glance "Swift backend does not support POST" [Critical,In progress] https://launchpad.net/bugs/713126 | 18:03 |
| uvirtbot | Launchpad bug 717431 in glance "Swift backend not taking into account new Swift account:user:key in URI" [High,Confirmed] https://launchpad.net/bugs/717431 | 18:03 |
| *** aryan__ has quit IRC | 18:03 | |
| *** pothos has quit IRC | 18:03 | |
| *** pothos_ has joined #openstack | 18:03 | |
| *** pothos_ is now known as pothos | 18:04 | |
| eday | jaypipes: ahh, perhaps I meant 'project' there for Nova | 18:04 |
| *** m_3 has joined #openstack | 18:04 | |
| eday | jaypipes: for swift, the account is always in the URI | 18:05 |
| jaypipes | eday: I think you may have meant "role", but it's not clear | 18:05 |
| *** mahadev has quit IRC | 18:05 | |
| eday | jaypipes: and swift (the service, no auth) only operates in the context of accounts... users is just something optional in an auth layer for swauth | 18:05 |
| eday | jaypipes: the users in swauth allow for different acls for the same account | 18:06 |
| eday | jaypipes: (and this is done via the "groups" | 18:06 |
| *** zul has joined #openstack | 18:07 | |
| *** grapex has quit IRC | 18:07 | |
| *** hvaldivia has quit IRC | 18:07 | |
| *** rcc has quit IRC | 18:07 | |
| *** londo_ has quit IRC | 18:07 | |
| eday | jaypipes: http://swift.openstack.org/overview_auth.html explains it pretty well | 18:07 |
| *** hvaldivia1 has joined #openstack | 18:07 | |
| *** westmaas has quit IRC | 18:07 | |
| *** jdarcy has quit IRC | 18:07 | |
| *** matclayton has quit IRC | 18:07 | |
| *** westmaas has joined #openstack | 18:07 | |
| *** ppetraki has quit IRC | 18:07 | |
| *** grapex has joined #openstack | 18:07 | |
| eday | jaypipes: reread email, "Nova currently has users and groups" should be "Nova currently has users and projects" | 18:07 |
| *** matclayton has joined #openstack | 18:08 | |
| *** matclayton has left #openstack | 18:08 | |
| *** jdarcy has joined #openstack | 18:08 | |
| jaypipes | eday: and for Swift, the account is NOT in the URI in the Swift docs. Could you point me to where you are getting that info? | 18:08 |
| *** ppetraki has joined #openstack | 18:08 | |
| eday | jaypipes: the account is set by the token auth server in the service URL | 18:08 |
| eday | jaypipes: for example, you speak to auth.openstack.org and get back a X-Storage-URL base of swift.openstack.org/v1/account_id | 18:09 |
| *** j05h has quit IRC | 18:11 | |
| *** flashn has quit IRC | 18:13 | |
| *** flashn has joined #openstack | 18:13 | |
| *** MarcMorata has quit IRC | 18:13 | |
| *** soren has quit IRC | 18:14 | |
| *** Seoman has joined #openstack | 18:16 | |
| eday | jaypipes: all operations then happen on top of that | 18:16 |
| eday | jaypipes: and http://docs.openstack.org/openstack-object-storage/developer/content/ch03s01.html#d5e307 does have it in there... | 18:16 |
| jaypipes | eday: You say "Nova currently has users and projects". Sure, but then you go ahead and talk about the concept of "roles" as well, and don't explain what they are. | 18:16 |
| jaypipes | eday: and please show me where the docs are for that in Swift. I can't find them. | 18:16 |
| *** kbringard has joined #openstack | 18:16 | |
| uvirtbot | New bug: #728597 in nova "OSAPI passes instance_id as unicode" [Low,In progress] https://launchpad.net/bugs/728597 | 18:16 |
| *** soren_ has joined #openstack | 18:16 | |
| jaypipes | eday: part of my frustration with Swift is the client tools. In Glance, we're using swift.common.client.Connection() and it's related methods like get_object(), put_container(), etc. NOWHERE is there a mention of account anywhere in swift.common.client.Connection()... | 18:16 |
| jaypipes | s/it's/its | 18:16 |
| *** joearnold has joined #openstack | 18:16 | |
| johnpur | eday|jaypipes: have you looked at: http: //plansthis.com/auth? | 18:16 |
| *** arun has joined #openstack | 18:17 | |
| *** arun has joined #openstack | 18:17 | |
| *** MarkAtwood has joined #openstack | 18:17 | |
| *** troytoman-away is now known as troytoman | 18:18 | |
| jaypipes | eday: sorry if I find all that documentation obtuse... it just doesn't explain how to *use* the Swift client to access multiple users and accounts... there's just no examples at all of using swift.common.client.Connection. | 18:18 |
| *** j05h has joined #openstack | 18:18 | |
| jaypipes | eday: have you used swift.common.client to actually programmatically connect to a Swift server? | 18:19 |
| johnpur | sheesh, how did that space get in there? http://plansthis.com/auth | 18:19 |
| eday | johnpur: yeah, I read it this morning | 18:22 |
| *** BK_man has quit IRC | 18:22 | |
| *** comstud has quit IRC | 18:22 | |
| *** cdbs has quit IRC | 18:22 | |
| *** dsockwell has quit IRC | 18:22 | |
| *** devcamcar has quit IRC | 18:22 | |
| *** soosfarm_ has quit IRC | 18:22 | |
| eday | jaypipes: I've not tried using swift.common.client, I'm just approaching it from HTTP directly | 18:22 |
| *** devcamcar has joined #openstack | 18:23 | |
| *** cdbs has joined #openstack | 18:23 | |
| *** cdbs has joined #openstack | 18:23 | |
| *** viirya has quit IRC | 18:23 | |
| *** zykes- has quit IRC | 18:23 | |
| *** widodh has quit IRC | 18:23 | |
| *** mfoemmel has quit IRC | 18:23 | |
| *** Daviey has quit IRC | 18:23 | |
| jaypipes | creiht: is swift.common.client even complete code? Should we not be using it? | 18:23 |
| *** mfoemmel has joined #openstack | 18:23 | |
| *** viirya has joined #openstack | 18:23 | |
| *** ianweller has quit IRC | 18:23 | |
| *** zykes- has joined #openstack | 18:23 | |
| *** widodh has joined #openstack | 18:24 | |
| *** soosfarm has joined #openstack | 18:24 | |
| *** dsockwell has joined #openstack | 18:24 | |
| *** jesse__ has joined #openstack | 18:24 | |
| *** jesse__ is now known as anotherjesse | 18:24 | |
| *** mahadev has joined #openstack | 18:25 | |
| *** jdarcy_ has joined #openstack | 18:25 | |
| *** thielmann__ has joined #openstack | 18:26 | |
| eday | jaypipes: it looks complete, the version/account is just passed as part of the url | 18:27 |
| *** ianweller has joined #openstack | 18:27 | |
| *** ianweller is now known as Guest3633 | 18:27 | |
| *** magglass2 has joined #openstack | 18:28 | |
| *** etet has quit IRC | 18:28 | |
| eday | jaypipes: for example: client.get_container('http://swift.com/v1/my_account', token, 'my_container') | 18:28 |
| *** MotoMilind has joined #openstack | 18:28 | |
| *** jdarcy_ has quit IRC | 18:28 | |
| *** allsystemsarego_ has joined #openstack | 18:28 | |
| *** irahgel1 has joined #openstack | 18:28 | |
| *** irahgel has quit IRC | 18:28 | |
| *** rackerhacker has quit IRC | 18:28 | |
| *** elasticdog has quit IRC | 18:28 | |
| *** jdarcy has quit IRC | 18:28 | |
| *** thielmann_ has quit IRC | 18:28 | |
| *** magglass1 has quit IRC | 18:28 | |
| *** ppetraki has quit IRC | 18:28 | |
| *** allsystemsarego has quit IRC | 18:28 | |
| *** gasbakid has quit IRC | 18:28 | |
| *** retr0h has joined #openstack | 18:29 | |
| *** retr0h has joined #openstack | 18:29 | |
| MotoMilind | Just wanted to confirm that Dashboard doesn't (yet) provide user management functionality, just role management. Thanks. | 18:29 |
| *** Guest3633 is now known as ianweller | 18:30 | |
| *** ianweller has joined #openstack | 18:30 | |
| anotherjesse | MotoMilind: dashboard supports user/project management in the django admin app | 18:30 |
| anotherjesse | MotoMilind: go to /admin | 18:30 |
| anotherjesse | jaypipes / eday this document / branch is a day of work of seeing what ideas from discussions fit nova / openstack | 18:30 |
| *** ppetraki has joined #openstack | 18:30 | |
| jaypipes | anotherjesse: hope to see it on the ML soon. | 18:31 |
| *** jdarcy has joined #openstack | 18:31 | |
| *** gasbakid has joined #openstack | 18:31 | |
| *** elasticdog has joined #openstack | 18:31 | |
| jaypipes | eday: hmm, ok. | 18:32 |
| *** rackerhacker has joined #openstack | 18:32 | |
| *** devcamca- has joined #openstack | 18:32 | |
| *** cdbs has quit IRC | 18:32 | |
| *** tr3buchet has quit IRC | 18:32 | |
| *** lool has quit IRC | 18:32 | |
| *** devcamcar has quit IRC | 18:32 | |
| *** arun has quit IRC | 18:32 | |
| *** notmyname has quit IRC | 18:32 | |
| *** yosh has quit IRC | 18:32 | |
| *** lool has joined #openstack | 18:32 | |
| *** arun has joined #openstack | 18:33 | |
| *** arun has joined #openstack | 18:33 | |
| *** notmyname has joined #openstack | 18:33 | |
| *** ChanServ sets mode: +v notmyname | 18:33 | |
| *** mfoemmel has quit IRC | 18:33 | |
| *** taihen has quit IRC | 18:33 | |
| *** taihen has joined #openstack | 18:33 | |
| *** cdbs has joined #openstack | 18:33 | |
| *** Daviey has joined #openstack | 18:33 | |
| eday | jaypipes: there are no openstack docs really that I can find on how it assume a token server passes back the base service URLs. That's in the rackspace version of the docs (rackspace.com/cloud api docs) | 18:33 |
| jaypipes | eday: so, is the plan to have all OS services use the Swift-style auth system? | 18:34 |
| *** tr3buchet has joined #openstack | 18:34 | |
| *** mfoemmel has joined #openstack | 18:34 | |
| *** yosh has joined #openstack | 18:34 | |
| tr3buchet | interesting | 18:34 |
| eday | jaypipes: there is no plan right now | 18:34 |
| jaypipes | anotherjesse: have you guys investigated using repoze.what instead of re-inventing a lot of those wheels? | 18:34 |
| eday | jaypipes: we still can't have a discussion since different folks are discussing it privately | 18:35 |
| jaypipes | eday: uhm, there seem to be a whole lot of plans going on behind the scenes off of the mailing lists... | 18:35 |
| eday | jaypipes: I was trying to drive discussioon on the ML the past couple days so we could have a plan, but unless folks are willing to participate, it's pointless, so I stopped | 18:36 |
| jaypipes | eday: yep. I'm catching up on emails since was in jury stuff ... | 18:36 |
| *** spectorclan has quit IRC | 18:36 | |
| *** jfluhmann has quit IRC | 18:39 | |
| *** jfluhmann has joined #openstack | 18:40 | |
| MotoMilind | Whoa, tight. I can go to /admin for user management, nifty. Thanks! | 18:40 |
| *** Adri2000 has quit IRC | 18:42 | |
| *** magglass2 has quit IRC | 18:48 | |
| *** rdw has quit IRC | 18:49 | |
| *** fysa has quit IRC | 18:50 | |
| *** magglass1 has joined #openstack | 18:50 | |
| *** magglass1 has joined #openstack | 18:50 | |
| *** rdw_ has joined #openstack | 18:50 | |
| *** nRy has joined #openstack | 18:52 | |
| *** Adri2000 has joined #openstack | 18:54 | |
| *** lool has quit IRC | 18:55 | |
| *** m_3 has quit IRC | 18:55 | |
| *** vvuksan has quit IRC | 18:56 | |
| *** paltman has quit IRC | 18:56 | |
| anotherjesse | jaypipes: the code is a proof of concept - using repoze.what (or another solution) is interesting | 18:57 |
| *** lool has joined #openstack | 18:57 | |
| jaypipes | anotherjesse: or, another question, why not use swift's auth design? | 18:58 |
| *** grapex has quit IRC | 18:58 | |
| anotherjesse | jaypipes: what parts of it? | 18:58 |
| anotherjesse | swift has internal account | 18:59 |
| anotherjesse | the point of this proposal is that projects shouldn't have accounts | 18:59 |
| eday | anotherjesse: projects in the nova sense, or projects as in other services? | 19:00 |
| nelson | question: should I be caching the auth token for $SOME_TIME, or should I be making a call to the auth server every time? | 19:00 |
| eday | anotherjesse: and if services don't use accounts, what do resources (server, swift object, ...) belong to? | 19:00 |
| jaypipes | anotherjesse: the design of Swift. in other words, having a middleware that inserts an X-Auth-Groups header in the environ and services down the pipe use that header's value... | 19:00 |
| nelson | The auth token times out after a period of time ... but if I ask the auth server for another one, it just gives me the same one. | 19:01 |
| *** zul_ has joined #openstack | 19:01 | |
| jaypipes | nelson: for Swift, I assume? | 19:01 |
| *** zul has quit IRC | 19:02 | |
| anotherjesse | jaypipes: I think we do that | 19:02 |
| anotherjesse | jaypipes: it is that we need to do more | 19:02 |
| anotherjesse | jaypipes: if we are pulling out the concept of users/projects from nova | 19:02 |
| anotherjesse | we need to add the ability to implement rbac and other policy questions | 19:02 |
| jaypipes | eday: in your ML email on auth, you say this about Swift: "All operations are done in the context of a user and account. A user | 19:03 |
| jaypipes | may not be a member of the account it's acting on since resources can specify ACLs, this is especially true for public resources (where user is undefined or anonymous)." Could you explain that? | 19:03 |
| anotherjesse | the IDM is in charge of roles (eg, groups and whatnot) | 19:03 |
| eday | anotherjesse: swauth already has that in there with "groups" and acl's stored per container | 19:03 |
| jaypipes | anotherjesse: and Swift does that by deferring to an external authz service that populates that X-Auth-Groups header... | 19:03 |
| anotherjesse | jaypipes: that is not authz - that is authn | 19:04 |
| eday | jaypipes: I was wrong about swift with users, one of the swift guys followed up and explained that's just one auth middleware (groups). swift is only accounts | 19:04 |
| jaypipes | anotherjesse: no, that is what is *returned* by the authn middleware and *used* by the authz middleware. | 19:04 |
| jaypipes | eday: k | 19:05 |
| nelson | jaypipes: yes, this is for swift. | 19:05 |
| nelson | jaypipes: there is a user/key pair, and account/token pair. | 19:06 |
| jaypipes | nelson: IIRC, you should just pass the token in every request after receiving it from the authentication service... the middleware should be checking the expiration of the token and updating the expiration date of that token... I think. :) | 19:06 |
| nelson | The auth server takes in user/key and gives back account/token. | 19:07 |
| jaypipes | ya | 19:07 |
| nelson | All the other services use account/token, although token is not needed if you've made a container public for reading or writing. | 19:07 |
| nelson | jaypipes: no, tokens definitely time out, and then you need to use a user/key to get a new one. | 19:08 |
| anotherjesse | jaypipes: do you know if there is a diagram kinda like http://twitpic.com/45qpvz/full | 19:08 |
| anotherjesse | for how swift works? | 19:09 |
| *** Xenith has quit IRC | 19:09 | |
| *** Xenith has joined #openstack | 19:09 | |
| nelson | anotherjesse: I haven't seen one. | 19:10 |
| jaypipes | anotherjesse: no idea. :( | 19:10 |
| nelson | but if you ask the auth server for another token, it will give you the same one. So ... my question is along the lines of "How often do I need to ask the auth server for a token". | 19:10 |
| notmyname | nelson: the auth server should return a TTL for the token | 19:10 |
| *** vvuksan has joined #openstack | 19:10 | |
| anotherjesse | jaypipes: http://paste.openstack.org/show/798/ is the text that built that (using an open source project called sdedit) | 19:11 |
| notmyname | you should be able to cache it for that time, but be sure to re-auth if you get a 401 from swift (eg the token could have been revoked) | 19:11 |
| jaypipes | nelson: so in swift, you can't have a user in >1 account? :( | 19:11 |
| anotherjesse | nelson: mike mayo says ~24 hours | 19:11 |
| nelson | Cuz surely it needs to give back auth tokens that change faster than their timeouts. | 19:11 |
| anotherjesse | jaypipes: cloud servers have the same model | 19:11 |
| jaypipes | annegentle: what model? | 19:12 |
| nelson | notmyname: it sends back a date, but it's now. | 19:12 |
| eday | anotherjesse: that is how swift works, minus the 'user' return when compute checks the token (only groups/account is returned) | 19:12 |
| anotherjesse | eday: we need user for audit logging | 19:12 |
| jaypipes | anotherjesse: that is an implementation detail. | 19:12 |
| notmyname | nelson: in general, you should be able to cache it indefinitely as long as you also handle getting 401s | 19:12 |
| anotherjesse | eday: cool - as we said in the proposal we wanted to be as similar to existing patterns as possible | 19:13 |
| eday | anotherjesse: well, "user" should just be the auth'd account | 19:13 |
| nelson | jaypipes: afaik, the mapping is from one user to one account | 19:13 |
| anotherjesse | eday: I disagree | 19:13 |
| jaypipes | nelson: well that's poop. | 19:13 |
| anotherjesse | eday: account != user | 19:13 |
| jaypipes | notmyname: is that true? in swift a user can only be in one account? | 19:13 |
| eday | anotherjesse: services should only deal with accounts IMO, which can be users, projects, ... | 19:13 |
| anotherjesse | eday: disagree - we need to know the "Account" and "who is doing this" (eg the user) | 19:14 |
| anotherjesse | the who is for more for usability / logging | 19:14 |
| nelson | notmyname: Hmmm.... that might be the simplest solution. Keep it until the server barfs on it. | 19:14 |
| eday | anotherjesse: but what if user is just another account? | 19:14 |
| anotherjesse | eday: it can be! | 19:14 |
| eday | anotherjesse: I agree the need for detailed logging | 19:14 |
| anotherjesse | eday: or it can be different if you need to | 19:14 |
| creiht | nelson: re caching, probably a good idea :) | 19:14 |
| anotherjesse | eg, a simple linux PAM identity backend would just say: account = user | 19:14 |
| *** Ryan_Lane|away is now known as Ryan_Lane | 19:14 | |
| notmyname | jaypipes: swift only knows about accounts, containers, and objects. the auth system is what maps the token/creds to a auth account/user/group/etc | 19:15 |
| *** rlucio has joined #openstack | 19:15 | |
| anotherjesse | a sophisticated AD/LDAP integration would use groups/projects as accounts & ldap users as user | 19:15 |
| * nelson puts his headspace back into the code. | 19:15 | |
| eday | anotherjesse: I just don't see the need for a abstract IDM to expose any other types than accounts and groups | 19:15 |
| jaypipes | notmyname: sorry, it's that last part of that sentence that has been so confusing... token/creds map to an account? or MORE than one account? and what is a "group" in that sense? | 19:15 |
| anotherjesse | eday: the IDM is something we plugin to | 19:16 |
| nelson | creiht: but before I do that ... I have a sweet patch to client.py. It lets you use put_object_chunked and send chunks as you get them. | 19:16 |
| anotherjesse | eday: not something we implement - eg - use someones ldap or pam or activedirectory ... | 19:16 |
| nelson | creiht: all will be published when I'm confident it's working. | 19:16 |
| eday | anotherjesse: i'm not saying the IDM won't deal with these, they of course need to with different backends, but the server (compute, swift, ...) should not be aware of those things | 19:16 |
| notmyname | jaypipes: it's all up to the auth implementation (which is probably talking in circles). swift will ask the auth server (via the middleware) if the given token is allowed for the swift account. | 19:17 |
| *** paltman has joined #openstack | 19:17 | |
| jaypipes | notmyname: but does the swift account allow multiple users? | 19:17 |
| nelson | jaypipes: of course. | 19:17 |
| Ryan_Lane | why would an account need multiple users? | 19:18 |
| *** joearnol_ has joined #openstack | 19:18 | |
| *** irahgel1 has left #openstack | 19:18 | |
| eday | anotherjesse: so, I think we do need another field being returned in the token check, saying which account ID is performing the action, I just wouldn't call that a user (since it's an account ID) | 19:18 |
| nelson | jaypipes: in principle there is no reason why multiple users couldn't be handed back the same X-Storage-URL. | 19:18 |
| Ryan_Lane | like in the case of a business account? | 19:18 |
| *** troytoman is now known as troytoman-away | 19:18 | |
| anotherjesse | Ryan_Lane: an account is a project in rackspace parlance | 19:18 |
| Ryan_Lane | ahhhh ok | 19:18 |
| *** ramkrsna has quit IRC | 19:18 | |
| *** joearnold has quit IRC | 19:18 | |
| jaypipes | nelson: sorry, above you just said that "The auth server takes in user/key and gives back account/token.". That implies the user and account are a one-to-one relationship, otherwise a user/key combo would return a *list* of accounts... | 19:19 |
| creiht | I think you guys are overcomplicating the auth server stuff :) | 19:19 |
| *** photron has joined #openstack | 19:19 | |
| nelson | jaypipes: no, user maps to account;not vice-versa. | 19:19 |
| jaypipes | creiht: well it's not like the stuff is particularly well documented. | 19:19 |
| anotherjesse | creiht: all we are saying is what you do with sending user + account | 19:19 |
| eday | anotherjesse: so I'm confused what the difference between 'user' and 'account' is with what the token server returns during the service check | 19:19 |
| jaypipes | nelson: wha? see this is what's driving me crazy... | 19:19 |
| eday | creiht: yes :) | 19:20 |
| anotherjesse | eday: I agre the names are confusing - we are using account / user in its rackspace cloud meaning | 19:20 |
| nelson | jaypipes: you use a user/key to get an account/token. | 19:20 |
| creiht | stepping back a bit | 19:20 |
| jaypipes | nelson: if I give the auth server a user and a key, and a user can be mapped to >1 account, then how the heck would giving the auth server a user and a key not return the *list* of accounts that the user belongs to? | 19:20 |
| creiht | swift, internally, only has a notion of accounts, containers for an account, and objects for a container | 19:20 |
| nelson | jaypipes: internally, swift consults the auth server to see if the token is valid for that account, but only the auth server cares about user/key. | 19:20 |
| anotherjesse | eday: in nova parlance: sending token to auth would return an opaque project & user identifier | 19:21 |
| eday | anotherjesse: rackspace doesn't have a concept of users though... | 19:21 |
| anotherjesse | eday: it does actually - it just isnt' documented well :) | 19:21 |
| nelson | jaypipes: you asked if one account could have multiple users; the answer is yes. | 19:21 |
| notmyname | jaypipes: for RAX the globablization project will do just that: return endpoints for each swift cluster (with perhaps a default one) | 19:21 |
| anotherjesse | eday: the rackspace model is: an account has many users. a user belongs to a single account | 19:21 |
| nelson | jaypipes: you also asked if one user could have multiple accounts; the answer is no. | 19:21 |
| eday | anotherjesse: hmm, we juse send account ID to IDM and get backa token.. the account:user thing is just a swauth thing, not a racskapce thing | 19:21 |
| eday | and I can't type | 19:22 |
| anotherjesse | it is also in ozone | 19:22 |
| nelson | creiht: could you read what I just wrote and verify that I'm not misleading jay? | 19:22 |
| creiht | anotherjesse: that actually isn't the rackspace model :) | 19:22 |
| jaypipes | fuck, I give up. I don't understand how an account can have multiple users, and a user can't have multiple accounts. that simply doesn't make sense. | 19:22 |
| creiht | everyhing is scrolling too fast | 19:22 |
| nelson | EVERYBODY STOP SO THAT CREIHT CAN CATCH UP. | 19:22 |
| eday | jaypipes: we're talking about different things here.. some do, some don't :) | 19:22 |
| nelson | :) | 19:22 |
| creiht | nelson: you are basically correct | 19:22 |
| creiht | lol | 19:22 |
| nelson | creiht: tnx. | 19:22 |
| jaypipes | creiht: sorry, how can an account have multiple users and a user not have multiple accounts? that doesn't make sense. | 19:23 |
| nelson | jaypipes: accounts and users are different things. It's just like you can have multiple coins in your pocket, but the could in your pocket all belong to you. | 19:23 |
| notmyname | jaypipes: swift has no concept of users. it only asks the auth service if a token can access the swift account. the auth service is free to map accounts to users in a many-to-many way | 19:23 |
| creiht | with swauth, you a user is assigned to a specific account, but with ACLS, you can give a user on one account access to a user on another count | 19:23 |
| nelson | jaypipes: what notmyname just said! | 19:23 |
| creiht | account | 19:23 |
| creiht | erm access to a container on another account | 19:23 |
| creiht | so trying again to step back a bit | 19:24 |
| jaypipes | god this stuff is so over complicated. | 19:24 |
| creiht | the current rackspace auth, only has a notion of accounts (and no users) | 19:24 |
| *** littleidea has joined #openstack | 19:25 | |
| nelson | devauth? | 19:25 |
| notmyname | nelson: no, the rackspace cloud auth | 19:26 |
| nelson | oh, so devauth introduces the idea of users? | 19:26 |
| creiht | nelson: correct | 19:26 |
| creiht | we use it internally for other legacy reasons | 19:26 |
| creiht | swift actually easily supports multiple auth systems at once | 19:26 |
| *** mgoldmann has quit IRC | 19:26 | |
| nelson | sure, because it only cares about account/tokens. | 19:27 |
| creiht | that is why we have the reseller prefeix | 19:27 |
| nelson | but how does it know which auth system to consult to verify the token is valid for the account? | 19:27 |
| creiht | prefix | 19:27 |
| nelson | is there a mapping from reseller prefix to auth system? | 19:27 |
| notmyname | yes | 19:27 |
| nelson | AHA! (light bulb moment) | 19:28 |
| creiht | nelson: more or less, but isn't totally obvious at first | 19:28 |
| jaypipes | nothing about any of it is "totally obvious". | 19:28 |
| creiht | We have a slightly modified auth middleware at RS that delegates which auth system to use based on the prefix | 19:28 |
| nelson | :) I feel your pain, jay, but stick with it. You'll get there. | 19:28 |
| creiht | lol | 19:28 |
| notmyname | to me, the complicated part is that for ALCs (authorization), the auth system gives swift an opaque string that can be used to compare against the swift account and/or container for permissions purposes | 19:30 |
| creiht | so lets get back to basics | 19:30 |
| notmyname | sure sure. that's really an optimization for scalability rather than a key feature | 19:30 |
| creiht | the whole purpose of auth is to determine if a given request can be made | 19:31 |
| redbo_ | fun-da-mentals | 19:31 |
| creiht | heh | 19:31 |
| jaypipes | creiht: yes, we know that. thx. | 19:31 |
| * creiht sighs | 19:31 | |
| anotherjesse | creiht: auth = authorization or authentication? | 19:32 |
| creiht | i usually mean both | 19:32 |
| creiht | but mostly authorization | 19:33 |
| anotherjesse | creiht: in swift do they have to be implemented at the same layer (eg authorization might be controlled by one system, authentication by another) | 19:33 |
| creiht | since authentication usually happens totally outside the end systems | 19:33 |
| jaypipes | creiht: an authn middleware should reply to a request that a user/key combination can enter a system and what groups, if any, the user belongs to. an authz middleware should return whether a user can complete a specific action. would you agree with this? | 19:33 |
| *** littleidea has quit IRC | 19:33 | |
| notmyname | the auth middleware in swift only does authorization. the authentication step happens out-of-band of swift | 19:34 |
| creiht | anotherjesse: a little bit of both | 19:34 |
| creiht | in the current implementation, authorization is based off of the idea of groups | 19:35 |
| creiht | the external auth system keeps track of the mapping of user<->groups | 19:35 |
| creiht | when swift validates a request with swauth, swauth returns the groups that the user belogs | 19:36 |
| jaypipes | creiht: isn't the "account" a "group"? | 19:36 |
| *** mdomsch has quit IRC | 19:36 | |
| anotherjesse | creiht: authentication happens outside of swift (as deployed in rackspace) and authorization is pluggable in swift - so while it has a rackspace specific authz process, it can be changed? | 19:36 |
| creiht | jaypipes: yes, the easiest groups are the account, and the account:user | 19:36 |
| jaypipes | creiht: wait! what is "the account:user"? | 19:37 |
| creiht | a unique string that identifies the user | 19:37 |
| anotherjesse | jaypipes: I think that is the string that authz parses? | 19:37 |
| creiht | since user by itself wouldn't be unique | 19:37 |
| anotherjesse | creiht: how does the rest of swift know the account? | 19:37 |
| anotherjesse | creiht: does it parse the "account:user" string | 19:37 |
| creiht | that way on another account, I could grant a container read acces to group "account:user" which would give that specific user access | 19:38 |
| eday | creiht: but account:user is only swauth, not swift service? | 19:38 |
| *** Vek has quit IRC | 19:38 | |
| jaypipes | sorry, I'm not understand what the "account:user" is and how a "unique string that identifies the user" is a "group"? | 19:38 |
| vishy | anotherjesse, creiht: we modeled the proposal after swauth in a lot of ways | 19:38 |
| vishy | so it would be easy to plugin | 19:39 |
| *** widodh has quit IRC | 19:39 | |
| redbo_ | yeah, account:user is just part of authentication, not authorization. | 19:39 |
| *** widodh has joined #openstack | 19:40 | |
| vishy | (changed the groups syntax slightly from "$account", "$account:$user" to "account:$account" and "user:$user" | 19:40 |
| creiht | eday: yes that is specific to swauth | 19:40 |
| eday | creiht: you may want to drop 'user' for now, and just talks about what swift service does/needs, not swauth (since that's a specific auth middleware) | 19:41 |
| *** redbo_ is now known as redbo | 19:41 | |
| * creiht backs up again | 19:41 | |
| creiht | :) | 19:41 |
| creiht | so given a request, the auth middelware provides two functions (1 for authentication and 1 for authorization) | 19:42 |
| vishy | since there is so much discussion going on right now | 19:42 |
| vishy | termie and jesse and I have been working on a proposal for authn authz | 19:42 |
| vishy | and we have sample code | 19:42 |
| notmyname | sample code!? but arguing is so much more productive! ;-) | 19:43 |
| vishy | ! | 19:43 |
| vishy | we're just polishing up the proposal, but I will provide a link in a sec | 19:43 |
| vishy | since everyone is debating about it at the moment :) | 19:43 |
| anotherjesse | notmyname: lolz | 19:44 |
| jaypipes | vishy: generally discussing something is done before proposing something, which is why we brought it up on the IRC channel. | 19:45 |
| vishy | jaypipes: we have discussed it a lot | 19:45 |
| jaypipes | vishy: where? | 19:45 |
| vishy | jaypipes: far too much imo. Bexar design summit mailing list, discussions amongst groups. | 19:46 |
| jaypipes | vishy: anyway, we're not debating, we're trying to understand from creiht how the Swift auth system actually works. | 19:46 |
| * creiht throws hands up and gets back to work | 19:46 | |
| creiht | gholt wrote the auth stuff anyways :) | 19:46 |
| eday | haha | 19:47 |
| jaypipes | vishy: all I remember discussing re: auth at the bexar design summit was jorge's short session on it, and there was little discussion in that session. | 19:47 |
| eday | vishy: as long as a service only deals with accounts and relationships/roles with other accounts (not users, projects, ...) I'm happy | 19:47 |
| jaypipes | vishy: re: mailing list and auth discussions, I see the thread eday started, but I see no responses from any Anso guys. | 19:48 |
| btorch | dos nova come with any nova.conf.sample somewhere ? the admin guide has some flags on ch03s03.html but that doesn't seem to be all possible flags | 19:48 |
| eday | vishy: and every resource (container, object, server, network, ...) should have one owner (an account) and has acls for other accounts to access it (the relationship/roles) | 19:49 |
| btorch | nova-compute --help (I guess all those flags right ?) | 19:49 |
| eday | vishy: I *think* thats what the etherpad with your proposal has, correct? | 19:49 |
| *** joearnol_ has quit IRC | 19:50 | |
| *** clauden_ has joined #openstack | 19:54 | |
| nelson | what's the difference between x-auth-token and x-storage-token | 19:54 |
| redbo | x-storage-token is eternally deprecated | 19:55 |
| nelson | curl is giving me back identical values for them. | 19:55 |
| vishy | http://plansthis.com/auth | 19:55 |
| vishy | (proposal) | 19:55 |
| nelson | redbo: so it hands back the same thing under two names so that programs looking for x-storage-token will still work? | 19:56 |
| *** jero has quit IRC | 19:56 | |
| redbo | yes | 19:56 |
| nelson | HTTPNoContent(headers={'x-auth-token': token, 'x-storage-token': token, 'x-storage-url': url}) | 19:57 |
| nelson | um, yeah, I'd guess so! | 19:57 |
| nelson | maybe this has already been done, but maybe get_auth should only be returning one of them? | 19:58 |
| *** Seoman has quit IRC | 19:59 | |
| *** h0cin has quit IRC | 20:05 | |
| *** anotherjesse has quit IRC | 20:06 | |
| *** jero has joined #openstack | 20:06 | |
| *** dfg has joined #openstack | 20:12 | |
| *** MarkAtwood has quit IRC | 20:12 | |
| *** blamar has joined #openstack | 20:15 | |
| *** jaypipes has quit IRC | 20:16 | |
| *** Ryan_Lane has quit IRC | 20:16 | |
| annegentle | btorch: I've found that sometimes not all the flags show up in the --help, but that's a good start. Here's another link too: | 20:17 |
| eday | vishy: in http://wiki.openstack.org/AuthnAuthz?action=AttachFile&do=view&target=auth.gif, can you define exactly what "user" and "account" are (and the types) on the return of "auth_middleware" | 20:17 |
| annegentle | btorch: http://wiki.openstack.org/FlagsGrouping | 20:18 |
| vishy | eday: the same user and account passed in to the idm | 20:18 |
| vishy | s/idm/authn | 20:19 |
| vishy | (when creating the token) | 20:19 |
| eday | vishy: ok | 20:19 |
| vishy | the account is important because it is how we define ownership | 20:20 |
| eday | vishy: so, IMO, user,pass (or perhaps it's an email, or access key, ...) should map to an account | 20:20 |
| vishy | the user is only important for logging in most cases | 20:20 |
| eday | so it should be user_account, but the raw thing the IDM took in | 20:20 |
| vishy | eday: I think that is a very non-userfriendly way to use it | 20:20 |
| vishy | I don't want to have to remember a different password for every account | 20:21 |
| vishy | and it doesn't allow special acls outside of the account | 20:21 |
| eday | vishy: why would you need to? | 20:21 |
| btorch | annegentle: cool thanks | 20:21 |
| vishy | say i want to give access to a given instance ot another project or another user | 20:21 |
| eday | then you use the account representing that user | 20:21 |
| vishy | you have problems with ownership in that case | 20:22 |
| vishy | say i give launch permission in my project to you | 20:22 |
| eday | vishy: it's a simple user:pass==another account in IDM. it's the same thing, just now a real account ID rather than a random blob that only the IDM should be aware of | 20:22 |
| vishy | if you login using your account, how does the system know that i should own the instance | 20:23 |
| redbo | nelson: yeah, we should probably remove that. It actually dates back to long before openstack. | 20:23 |
| vishy | ownership simplifies authz in a number of ways | 20:23 |
| vishy | as well as data retrieval | 20:23 |
| eday | vishy: because I would POST to /version/your_account/servers, not /version/my_account/servers | 20:24 |
| eday | vishy: I'm all for ownership, and agree it simplifies things | 20:24 |
| vishy | eday: so you are passing account, you're just doing it in the url | 20:24 |
| *** ctennis_ has quit IRC | 20:24 | |
| eday | vishy: yes, but I'm passing a IDM account ID, not a random user string | 20:24 |
| vishy | i don't want to require a resty url interface to use the system, so account must be included in the IDM | 20:24 |
| *** mfoemmel has quit IRC | 20:24 | |
| *** Ryan_Lane has joined #openstack | 20:25 | |
| eday | vishy: I'm not saying "user" shouldn't be in there, I'm saying "user" should be another account | 20:25 |
| *** mfoemmel has joined #openstack | 20:25 | |
| vishy | eday: I gotta go get some food, perahps this would be easier over skype when i get back? | 20:25 |
| eday | vishy: and IDM does user:pass<->account_id translation, and services *only* deal with account IDs | 20:26 |
| eday | vishy: sure | 20:26 |
| *** zul_ is now known as zul | 20:26 | |
| *** gasbakid has quit IRC | 20:33 | |
| *** adiantum has quit IRC | 20:42 | |
| *** greenisus has joined #openstack | 20:44 | |
| *** Adri2000 has quit IRC | 20:47 | |
| *** miclorb_ has joined #openstack | 20:55 | |
| vishy | eday: bakc | 20:59 |
| vishy | s/bakc/back | 20:59 |
| *** allsystemsarego_ has quit IRC | 21:00 | |
| *** littleidea has joined #openstack | 21:00 | |
| creiht | I have at least found a feature in the lp slowness in that if I click on a link then realize that I clicked on the wrong link, I usually have time to click on the correct link before the wrong page shows up :) | 21:00 |
| eday | creiht: haha, I use that "feature" too :) | 21:03 |
| eday | vishy: I'm ready whenever | 21:03 |
| *** zul has quit IRC | 21:04 | |
| *** jesse_ has joined #openstack | 21:04 | |
| *** Adri2000 has joined #openstack | 21:05 | |
| *** mahadev has quit IRC | 21:05 | |
| *** jesse_ is now known as anotherjesse | 21:08 | |
| *** littleidea has quit IRC | 21:10 | |
| *** blueadept has quit IRC | 21:11 | |
| *** dprince has quit IRC | 21:12 | |
| *** jero has quit IRC | 21:14 | |
| *** jero has joined #openstack | 21:15 | |
| openstackhudson | Project swift build #210: SUCCESS in 29 sec: http://hudson.openstack.org/job/swift/210/ | 21:16 |
| openstackhudson | * Tarmac: Refactor proxy for concurrency and code reuse | 21:16 |
| openstackhudson | * Tarmac: Skip the swift3 middleware tests if boto is not installed (since it is an optional component) | 21:16 |
| *** joearnold has joined #openstack | 21:17 | |
| *** mahadev has joined #openstack | 21:18 | |
| johnpur | folks, there is a new blog post talking about OpenStack governance updates here: http://www.openstack.org/blog/2011/03/openstack-governance-update/ | 21:26 |
| johnpur | details on the OpenStack governance model are here: http://wiki.openstack.org/Governance/Model | 21:27 |
| *** MarkAtwood has joined #openstack | 21:30 | |
| *** joearnold has quit IRC | 21:30 | |
| devcamca- | soren, mtaylor, jaypipes: i could use an assist fixing hudson for openstack-dashboard | 21:31 |
| *** Vek has joined #openstack | 21:34 | |
| devcamca- | http://paste.openstack.org/show/799/ | 21:34 |
| *** ctennis has joined #openstack | 21:36 | |
| *** greenisus has quit IRC | 21:38 | |
| *** littleidea has joined #openstack | 21:38 | |
| devcamca- | let me rephrase | 21:41 |
| devcamca- | is anyone on that has access to our hudson environments? | 21:41 |
| devcamca- | been trying to get this fixed for 4 days now | 21:41 |
| *** littleidea has quit IRC | 21:42 | |
| btorch | ok, what is this --key supposed to be when trying to use euca-run-instances ? I have tried to use novacreds and novacreds.zip | 21:42 |
| kpepple | vishy: thank you for the auth.gif picture. question: does identity:Service in your diagram map to something like LDAP or is it a openstack service that mediates/proxies/relays to it ? or does it matter ... | 21:44 |
| annegentle | btorch: it's from the euca-add-keypair command, not from the novacreds collection of creds | 21:45 |
| btorch | hehe did I miss that step or that is not mentioned on the installation/configuration ? I don't remember seeing it but I'll double check it | 21:46 |
| annegentle | btorch: yeah I'm looking also, I think it's missing from the manual install instructs | 21:47 |
| mtaylor | devcamca-: yeah - I keep meaning to look at that... | 21:47 |
| annegentle | btorch: well it's really a config step | 21:47 |
| * mtaylor bitches: this is the reason I hate venv ... | 21:47 | |
| annegentle | btorch: yeah it's missing from http://docs.openstack.org/openstack-compute/admin/content/ch05s01.html, I'll add it | 21:48 |
| annegentle | btorch: good catch, thanks for asking | 21:48 |
| johnpur | yay, Monty is here! | 21:48 |
| mtaylor | devcamca-: ok. sadly, given the time you've been waiting, that seems to be an easy fix - can you try again please? | 21:48 |
| devcamca- | mtaylor: its odd, i dont know what changed | 21:49 |
| devcamca- | yep trying now | 21:49 |
| mtaylor | devcamca-: it was the change from hudson to jenkins - there was a /tmp/distribute-*tar.gz in /tmp owned by the hudson user | 21:49 |
| devcamca- | ah hah | 21:49 |
| devcamca- | makes sense | 21:49 |
| *** devcamca- is now known as devcamcar | 21:49 | |
| mtaylor | devcamca-: of course, the fact that the build was leaving stuff around in tmp is a bit fail - but not really your fault | 21:49 |
| mtaylor | there are a bazillion leftover tmpdirs in there ... | 21:50 |
| devcamcar | mtaylor: odd | 21:50 |
| devcamcar | mtaylor: why is that? | 21:50 |
| mtaylor | devcamcar: some of the openstack test suites create stuff in tempdirs and then don't cleanup after themselves | 21:50 |
| btorch | annegentle: cool thanks | 21:50 |
| * mtaylor thought jaypipes was fixing that... | 21:50 | |
| devcamcar | mtaylor: ah, so nothing to do with this repo | 21:51 |
| anticw | swift q, in testing bexar i see a lot of .db.pending databases accumulating over time | 21:53 |
| anticw | what process(es) would normally be responsible for dealing with that | 21:53 |
| anticw | (the pending all seem to have 0 filesize) | 21:54 |
| *** anotherjesse has quit IRC | 21:59 | |
| *** bcwaldon has joined #openstack | 22:03 | |
| *** Ephur has quit IRC | 22:05 | |
| *** photron has quit IRC | 22:08 | |
| devcamcar | mtaylor: its been about 15 mins since I clicked approve, and no word from hudson | 22:11 |
| devcamcar | mtaylor: hm, looks like it merged but no notification was sent out. maybe i'm just being impatient | 22:14 |
| *** hazmat has joined #openstack | 22:14 | |
| mtaylor | devcamcar: sometimes the emails take a bit | 22:15 |
| devcamcar | mtaylor: yea i think everything is happy again | 22:15 |
| mtaylor | devcamcar: yay! | 22:16 |
| openstackhudson | Project swift build #211: SUCCESS in 29 sec: http://hudson.openstack.org/job/swift/211/ | 22:16 |
| openstackhudson | * Tarmac: Add a "Getting Swift" section to the getting started doc | 22:16 |
| openstackhudson | * Tarmac: Ability to fasttrack auditing of zero byte files. | 22:16 |
| devcamcar | mtaylor: thanks! | 22:17 |
| *** Hello92 has joined #openstack | 22:18 | |
| *** bcwaldon has quit IRC | 22:22 | |
| Hello92 | So, say I have 8 nodes that are each swift + compute + ceph nodes all in one (is that a good idea?), on a gigabit lan, and I want to start the 8th virtual machine on the "cloud", how long after I give the command can I expect the vm up and running? Is it a matter of several seconds? Minutes? Milliseconds? | 22:22 |
| *** Vek has quit IRC | 22:26 | |
| *** adiantum has joined #openstack | 22:28 | |
| *** kbringard has quit IRC | 22:29 | |
| *** joearnold has joined #openstack | 22:29 | |
| *** aliguori has quit IRC | 22:30 | |
| *** joearnold has joined #openstack | 22:31 | |
| openstackhudson | Project swift build #212: SUCCESS in 29 sec: http://hudson.openstack.org/job/swift/212/ | 22:31 |
| openstackhudson | Tarmac: update functional tests configuration | 22:31 |
| *** m_3 has joined #openstack | 22:33 | |
| *** hstrange has joined #openstack | 22:36 | |
| *** hstrange has quit IRC | 22:37 | |
| *** greenisus has joined #openstack | 22:41 | |
| *** greenisus has quit IRC | 22:47 | |
| *** Ryan_Lane has quit IRC | 22:47 | |
| occ | I'm trying to set up a multi node compute cluster, where cloud controller would not run any VMs, and compute nodes would not run anything relating to the API or cloud controller. I couldn't find which services should be on which servers (documentation simply says "Install all the nova- packages and dependencies"). Is there a document explaining the roles of these services? | 22:48 |
| btorch | what's up with this virbr0 ? I'm using the flat manage and I think the iptables might be kind of messed up | 22:48 |
| btorch | I have created the br100 with eth2 as part of it ... br100 is configured with 10.0.0.10 within the 10.0.0.0/8 network | 22:49 |
| annegentle | occ: ok, this is an overview: http://docs.openstack.org/openstack-compute/admin/content/ch05.html and are you quoting from http://docs.openstack.org/openstack-compute/admin/content/ch05.html? | 22:52 |
| occ | annegentle: yep, but it's not descriptive enough. Like- do I need 1 nova-network instance on the cloud controller, or do I need to install nova-network on every node? same for scheduler and volume | 22:53 |
| occ | annegentle: sorry, I was quoting from http://docs.openstack.org/openstack-compute/admin/content/ch03s02.html#d5e239 | 22:55 |
| annegentle | occ: yeah those just walk through a 2-node install | 22:56 |
| uvirtbot | New bug: #728750 in openstack-dashboard "UserData kb page not found" [Undecided,New] https://launchpad.net/bugs/728750 | 22:56 |
| occ | hmm, is there anywhere I could get help from? | 22:57 |
| *** adiantum has quit IRC | 23:00 | |
| *** dovetail1an has joined #openstack | 23:00 | |
| *** dovetail1an has joined #openstack | 23:01 | |
| *** dovetaildan has quit IRC | 23:02 | |
| annegentle | occ: I think you're basically wanting the same config as dubsquared has done - see http://www.dubsquared.com/?p=43 and http://www.dubsquared.com/?p=100. You should just need nova-compute on the compute node. | 23:05 |
| *** adiantum has joined #openstack | 23:08 | |
| annegentle | occ: he's not on IRC right now, but you could ask him questions on the blog post | 23:09 |
| hvaldivia1 | Hi everybody. | 23:10 |
| hvaldivia1 | FlatDHCPManager uses dnsmaq for the ips, right? | 23:10 |
| rlucio | can anyone comment on whether or not any xen based hypervisor, say on ubuntu or rhel, is supported? or is it just xenserver | 23:10 |
| rlucio | * for bexar | 23:11 |
| hvaldivia1 | Would it be possible to set dnsmaq as a dchp-proxy or relay to get the IPs from a DHCP Server in other machine? | 23:11 |
| *** ppetraki has quit IRC | 23:13 | |
| hvaldivia1 | Does anyone have an idea? | 23:15 |
| *** hazmat has quit IRC | 23:15 | |
| vvuksan | hvaldivia1: I doubt it | 23:16 |
| vvuksan | hvaldivia1: because nova-network will add MAC address and IP to dnsmasq static mapping | 23:16 |
| *** maplebed has joined #openstack | 23:19 | |
| hvaldivia1 | nova-network uses nova.conf ( fixed_range, etc ) —> IP and MAC —> dnsmasq —> VM? | 23:20 |
| hvaldivia1 | Is that right? | 23:20 |
| occ | annegentle: thanks | 23:20 |
| hvaldivia1 | Ok, it would be hard to modify nova-network to support dhcp-proxy | 23:21 |
| *** maplebed has quit IRC | 23:22 | |
| *** maplebed has joined #openstack | 23:22 | |
| *** littleidea has joined #openstack | 23:26 | |
| *** mray has quit IRC | 23:28 | |
| *** freeflyi1g has joined #openstack | 23:32 | |
| *** hvaldivia1 has quit IRC | 23:33 | |
| *** adiantum has quit IRC | 23:35 | |
| *** jero has quit IRC | 23:35 | |
| *** freeflying has quit IRC | 23:35 | |
| *** londo_ has joined #openstack | 23:38 | |
| *** slyphon has quit IRC | 23:39 | |
| *** MarkAtwood has quit IRC | 23:39 | |
| *** adiantum has joined #openstack | 23:43 | |
| *** MotoMilind has quit IRC | 23:47 | |
| *** joearnold has quit IRC | 23:56 | |
| *** adiantum has quit IRC | 23:56 | |
| *** littleidea has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
