| *** reed_ has quit IRC | 00:01 | |
| heckj | anotherjesse: https://review.openstack.org/#change,5054 <-- what's the background on this one? | 00:04 |
|---|---|---|
| anotherjesse | heckj: it wasn't removed from the header docs | 00:04 |
| heckj | ah | 00:04 |
| anotherjesse | grep through the code | 00:04 |
| anotherjesse | no meantion of authorization | 00:04 |
| anotherjesse | heckj: we should get dolph to review as well | 00:04 |
| heckj | dolphm: ^^ (yeah) | 00:05 |
| *** oneiroi has quit IRC | 00:05 | |
| *** oneiroi has joined #openstack-dev | 00:12 | |
| *** zykes- has quit IRC | 00:12 | |
| *** michaelchapman has quit IRC | 00:14 | |
| *** zykes- has joined #openstack-dev | 00:15 | |
| *** michaelchapman has joined #openstack-dev | 00:16 | |
| *** reed_ has joined #openstack-dev | 00:16 | |
| *** lloydde has joined #openstack-dev | 00:19 | |
| *** anotherjesse is now known as anotherjesse_zz | 00:20 | |
| *** lloydde has quit IRC | 00:22 | |
| *** flaviamissi has joined #openstack-dev | 00:24 | |
| *** lloydde has joined #openstack-dev | 00:24 | |
| *** dtroyer has quit IRC | 00:25 | |
| *** flaviamissi_ has joined #openstack-dev | 00:28 | |
| *** flaviamissi has quit IRC | 00:29 | |
| *** hub_cap has joined #openstack-dev | 00:29 | |
| *** flaviamissi_ is now known as flaviamissi | 00:32 | |
| *** flaviamissi has quit IRC | 00:34 | |
| *** flaviamissi has joined #openstack-dev | 00:34 | |
| *** lloydde has quit IRC | 00:37 | |
| *** lloydde has joined #openstack-dev | 00:45 | |
| *** littleidea has quit IRC | 00:50 | |
| *** littleidea has joined #openstack-dev | 00:57 | |
| *** yamahata__ has quit IRC | 00:58 | |
| *** yamahata__ has joined #openstack-dev | 00:59 | |
| *** sniperd has joined #openstack-dev | 01:00 | |
| hub_cap | hey if there are some keystone devs around, plz review. fixing as per the api spec. https://review.openstack.org/5055 | 01:00 |
| *** littleidea has quit IRC | 01:01 | |
| mnewby | Any keystone devs here? | 01:02 |
| *** sandywalsh has quit IRC | 01:02 | |
| *** littleidea has joined #openstack-dev | 01:03 | |
| mnewby | Or rather, anybody that understands the cfg module? | 01:03 |
| *** littleidea has quit IRC | 01:05 | |
| hub_cap | mnewby: shoot, i might or might not be able to help… im usually not helpful ;) | 01:05 |
| mnewby | I've found a bug in devstack related to a keystone config change, but am wondering if the fix needs to happen in devstack or keystone. | 01:05 |
| mnewby | https://github.com/openstack/keystone/commit/d0429ea9b8849f99aa170cd09aef7776e2651dbf#etc/keystone.conf | 01:05 |
| mnewby | This change introduced a configurable bind_host. Devstack's version of keystonel.conf doesn't include bind_host, so the stack run fails to start keystone. | 01:06 |
| hub_cap | well if the default keystone.conf has a bind_host id say its devstacks problem, right? | 01:07 |
| hub_cap | by default keystone.conf i mean the one in keystones repo | 01:07 |
| mnewby | I'm looking at keystone.config and it registers bind_host with a default, but in the keystone-all script the options don't reflect a value for bind_host unless it is explicitly defined. | 01:07 |
| hub_cap | sry for the confusing wording :P | 01:07 |
| mnewby | Is that by design? | 01:07 |
| mnewby | I'm a bit confused as to why a default would be specified in keystone.config but that default would not be reflected in the loaded options. | 01:08 |
| mnewby | Does that question make sense? | 01:08 |
| hub_cap | def, give me a min to look about | 01:08 |
| mnewby | ok | 01:09 |
| hub_cap | well keystone_all loads the default config file in its repo, and then prints it all out. do u not see that bind_host printed when keystone starts? | 01:10 |
| *** pixelbeat has quit IRC | 01:10 | |
| hub_cap | let me reclone my devstack repo and see what happens | 01:10 |
| mnewby | hub_cap: I see the output, and the output does include bind_host. | 01:11 |
| mnewby | But the value of the options variable does not reflect the defaults set in CONF | 01:11 |
| mnewby | It looks like the options are being loaded without regard to the default in CONF | 01:11 |
| mnewby | I'm guessing that is a bug in keystone - otherwise, why would the logged option values not reflect the options that are to be used? | 01:12 |
| hub_cap | so i do see the config.py in keystone is setting that and defaulting it | 01:12 |
| hub_cap | register_str('bind_host', default='0.0.0.0') | 01:12 |
| mnewby | Check the keystone-all sscript | 01:13 |
| hub_cap | ya it grabs config.CONF | 01:13 |
| mnewby | On line 61, the CONF values are being output. | 01:13 |
| mnewby | But on line 63, the options are being loaded directly from the file. | 01:13 |
| hub_cap | ya for sure thats why i asked u if it was printed :P | 01:13 |
| hub_cap | k sec | 01:13 |
| hub_cap | but even so, it should be printing it on 61 right? | 01:13 |
| mnewby | So options is not being updated with the CONF defaults. | 01:13 |
| mnewby | Yes, it is printing on 61. But what is being printed differs from what is actually used. | 01:14 |
| hub_cap | ah ok so it _is_ printed | 01:14 |
| hub_cap | the config stuff that im doing w/ reddwarf is def different from this. my devstack install is _still_ rebooting ;) | 01:15 |
| hub_cap | err recloning | 01:15 |
| hub_cap | id agree that it migth not be a devstack issue if keystone, w/o devstack, fails starting cuz of this | 01:15 |
| mnewby | It is an easy thing to repro - try to start keystone without a 'bind_host' entry in keystone.conf. | 01:16 |
| *** sandywalsh has joined #openstack-dev | 01:16 | |
| mnewby | The logged output will show the default value for bind_host, but the startup will fail with a KeyError for bind_host. | 01:16 |
| hub_cap | exactly, ill do that w/o devstack once i get it all recloned | 01:16 |
| hub_cap | note line 68 in it, try printing out options['bind_host'] and see if its blank | 01:17 |
| *** flaviamissi has quit IRC | 01:18 | |
| mnewby | If bind_host is not in keystone.conf, the failure will be KeyError on line 68 | 01:18 |
| hub_cap | yup | 01:18 |
| hub_cap | is that the case? | 01:18 |
| mnewby | Yes | 01:18 |
| hub_cap | ya seems like a keystone bug then for sure | 01:19 |
| mnewby | Ok | 01:19 |
| hub_cap | yup failing for me too, w/ a fresh devstack | 01:19 |
| mnewby | Does any other openstack project use cfg to set defaults? | 01:19 |
| mnewby | I'm wondering how the explicit config should get merged with the default config. | 01:20 |
| hub_cap | oh wait up | 01:20 |
| hub_cap | this is _not_ a keystone issue | 01:20 |
| mnewby | ? | 01:20 |
| hub_cap | CONF.config_file[0] | 01:21 |
| mnewby | The devstack version of keystone.conf is missing bind_host | 01:21 |
| hub_cap | right thats the issue like u stated earlier | 01:21 |
| mnewby | But keystone is still printing out options that don't reflect what's used. | 01:21 |
| hub_cap | its really a dual issue i think | 01:21 |
| hub_cap | right, it uses defaults, but then wipes them out, just like you said | 01:21 |
| mnewby | It doesn't wipe them, it just doesn't apply them to the loaded config. | 01:22 |
| hub_cap | id say keystone is at fault, but devstack should also be notified of the issue to add it | 01:22 |
| mnewby | It only prints them out. | 01:22 |
| mnewby | Ok | 01:22 |
| mnewby | I can add a change to devstack easily enough. | 01:22 |
| mnewby | Or rather, submit a change. | 01:22 |
| mnewby | The keystone issue is a bit more problematic. | 01:22 |
| *** jdg has quit IRC | 01:23 | |
| hub_cap | right | 01:23 |
| *** dtroyer has joined #openstack-dev | 01:27 | |
| hub_cap | ya looking at the pastedeploy src there is no easy way to munge them together | 01:29 |
| hub_cap | the new openstackcommon stuff simplifies this so they may need to move some of the wsgi paste stuff to use that | 01:29 |
| mnewby | Ah, hahah | 01:30 |
| mnewby | heckj already committed a fix. | 01:30 |
| heckj | mnewby: and there's an outstanding bug to made the defaults work correctly, which I think will flow when Mark's openstack common config reviews get through and merged. | 01:31 |
| mnewby | Ah, ok. | 01:31 |
| mnewby | Good to know it is being looked at then. | 01:32 |
| hub_cap | haha nice work heckj | 01:32 |
| hub_cap | i can sees fix? | 01:32 |
| *** tserong has quit IRC | 01:34 | |
| mnewby | arg | 01:34 |
| mnewby | my dev box was using the cloudbuilder's devstack | 01:34 |
| *** lloydde has quit IRC | 01:34 | |
| *** jaypipes has joined #openstack-dev | 01:35 | |
| hub_cap | oh fix in devstack | 01:35 |
| hub_cap | i was hoping to see a keystone fix | 01:35 |
| mnewby | heckj: When did devstack switch from cloudbuilders to openstack-dev? | 01:35 |
| *** tserong has joined #openstack-dev | 01:35 | |
| heckj | mnewby: some time back, actually. When we started gating commits to projects on it's basic integration testing | 01:36 |
| mnewby | heckj: Ah, ok. I'd just been blithely git pulling from cloudbuilders for the past couple of weeks. It's a wonder more didn't break! | 01:36 |
| *** Kiall has quit IRC | 01:37 | |
| heckj | they keep it sync'd pretty close. | 01:37 |
| mnewby | That explains it. | 01:37 |
| *** tserong has quit IRC | 01:39 | |
| *** Kiall has joined #openstack-dev | 01:40 | |
| *** tserong has joined #openstack-dev | 01:40 | |
| *** cp16net has joined #openstack-dev | 01:40 | |
| *** tserong has quit IRC | 01:41 | |
| *** tserong has joined #openstack-dev | 01:41 | |
| mnewby | heckj: While I'm in the business of distracting you, could I ask a question about auth_token and swift_auth? | 01:42 |
| heckj | mnewby: I'll do my best to answer | 01:42 |
| heckj | (swift_auth is not my strong point) | 01:42 |
| mnewby | I used them as the basis for some swift auth stuff I did here at Internap, and have written some tests for swift_auth. If i wanted to submit them, would I need a launchpad bug? | 01:43 |
| mnewby | Also, I created a simple Identity class that would serialize to and from headers to simplify the code in both auth_token and swift_auth. Does this sound like something worth contributing? | 01:44 |
| heckj | It's generally best to have either a bug or a blueprint for code that you're submitting. I'm still getting our blueprints (which typically represent new features or functionality) reset to lead into the folsom design summit, so right now there's a pile of bugs that should really be blueprints (for Keystone) | 01:44 |
| mnewby | Ok, good to know. | 01:45 |
| mnewby | So a change that adds tests could target a bug complaining of missing tests? | 01:46 |
| heckj | mnewby: more tests is ALWAYS good :-) | 01:47 |
| mnewby | ok :) | 01:47 |
| heckj | mnewby: the class sounds useful, but I'd want to generally review it with the other core devs | 01:48 |
| hub_cap | tests ftw! | 01:48 |
| mnewby | So a blueprint for the class addition. | 01:49 |
| notmyname | mnewby: can you add some simple tox docs to the swift docs? how to run, etc. I think that's the only outstanding issue. something like "to test with tox, install like this and run like this". I think adding it to the http://swift.openstack.org/development_guidelines.html page would be good | 01:49 |
| mnewby | notmyname: can do | 01:50 |
| notmyname | mnewby: thanks :-) | 01:50 |
| *** littleidea has joined #openstack-dev | 01:51 | |
| *** heckj has quit IRC | 01:57 | |
| *** gyee has quit IRC | 02:09 | |
| *** anotherjesse_zz is now known as anotherjesse | 02:10 | |
| *** anotherjesse is now known as anotherjesse_zz | 02:12 | |
| *** anotherjesse_zz is now known as anotherjesse | 02:16 | |
| *** reed_ has quit IRC | 02:16 | |
| *** anotherjesse is now known as anotherjesse_zz | 02:18 | |
| *** mszilagyi has quit IRC | 02:18 | |
| *** johnpostlethwait has joined #openstack-dev | 02:18 | |
| *** johnpostlethwait has left #openstack-dev | 02:19 | |
| *** johnpostlethwait has joined #openstack-dev | 02:19 | |
| *** anotherjesse_zz is now known as anotherjesse | 02:19 | |
| *** johnpostlethwait has quit IRC | 02:20 | |
| *** anotherjesse is now known as anotherjesse_zz | 02:20 | |
| *** sniperd has quit IRC | 02:24 | |
| *** vincentricci has quit IRC | 02:28 | |
| *** adjohn has quit IRC | 02:28 | |
| *** dalang has quit IRC | 02:30 | |
| *** hugokuo has joined #openstack-dev | 02:35 | |
| rmk | So, when I try a connection to the wsproxy using Diablo, I initially get an invalid token response before it succeeds | 02:37 |
| rmk | As in, I need to try again | 02:37 |
| *** danwent has quit IRC | 02:43 | |
| *** vizsla has joined #openstack-dev | 02:51 | |
| *** Ryan_Lane has quit IRC | 02:51 | |
| *** rkukura has joined #openstack-dev | 02:56 | |
| *** maplebed_ has joined #openstack-dev | 02:59 | |
| *** maplebed_ is now known as Guest32353 | 02:59 | |
| *** maplebed has quit IRC | 03:01 | |
| *** novas0x2a|laptop has quit IRC | 03:01 | |
| *** Guest32353 is now known as maplebed | 03:04 | |
| *** maplebed has joined #openstack-dev | 03:04 | |
| *** ayoung has quit IRC | 03:07 | |
| *** spiffxp has quit IRC | 03:09 | |
| *** madhav has joined #openstack-dev | 03:09 | |
| *** madhav has left #openstack-dev | 03:10 | |
| *** rkukura has quit IRC | 03:13 | |
| *** mszilagyi has joined #openstack-dev | 03:14 | |
| *** rkukura has joined #openstack-dev | 03:17 | |
| *** martine has joined #openstack-dev | 03:21 | |
| *** sniperd has joined #openstack-dev | 03:39 | |
| rmk | Connection to vnc works every other time. | 03:42 |
| *** anotherjesse_zz is now known as anotherjesse | 03:47 | |
| *** rkukura has quit IRC | 03:55 | |
| *** jdurgin has quit IRC | 04:01 | |
| *** maplebed has quit IRC | 04:09 | |
| andrewbogott | Are any nova developers still up? | 04:09 |
| anotherjesse | mnewby: we have a script that keeps cloudbuilder's repo up-to-date & doc'd - but the github ssh issue caused our jenkins job to break | 04:10 |
| anotherjesse | fixed and updated | 04:10 |
| mnewby | anotherjesse: ah, makes sense | 04:11 |
| *** vincentricci has joined #openstack-dev | 04:11 | |
| andrewbogott | I'm confused because in Horizon I see that I have four projects; but in my code I'm calling project_get_all and getting an empty list. Why would they disagree? | 04:12 |
| *** vincentricci has left #openstack-dev | 04:15 | |
| anotherjesse | andrewbogott: diablo or essex? | 04:15 |
| andrewbogott | essex. | 04:16 |
| anotherjesse | keystone or deprecated auth? | 04:16 |
| andrewbogott | I've rebased in the last couple of days. | 04:16 |
| andrewbogott | Whatever devstack does... keystone, I think. | 04:16 |
| anotherjesse | andrewbogott: the project_get_all code is nova code that goes against deprecated auth | 04:16 |
| andrewbogott | Oh... does that mean that the whole db api is deprecated? | 04:17 |
| anotherjesse | yeah - all users exist in keystone | 04:17 |
| anotherjesse | andrewbogott: to get your list of projects you can use keystone's get_tenants_for_user | 04:17 |
| anotherjesse | with a token you can call /v2.0/tenants (which is what the dash is doing) | 04:17 |
| andrewbogott | Ok... this almost makes sense. | 04:18 |
| andrewbogott | Is 'tenant' a synonym for project? I thought a tenant was... something else. | 04:18 |
| anotherjesse | swift's accounts / nova's project = keystone's tenants | 04:19 |
| anotherjesse | although the horizon folks don't like the word tenant so they show project in the dropdown | 04:19 |
| andrewbogott | iirc horizon uses both words in different contexts. | 04:20 |
| anotherjesse | yeah, the switch to project was recent … and didn't ripple through | 04:20 |
| andrewbogott | So, users and projects (aka tenants) are managed by keystone now... | 04:21 |
| andrewbogott | So, to manage the permissions of an instance, the nova db holds keys into a keystone db? | 04:21 |
| andrewbogott | Or do keystone tenants hold a list of instances? | 04:22 |
| andrewbogott | (does this question make sense, even?) | 04:22 |
| *** thingee has joined #openstack-dev | 04:23 | |
| anotherjesse | andrewbogott: resources are owned by their service - nova knows about servers, volumes, ips, ... | 04:24 |
| anotherjesse | keystone knows about uesrs/tenants/... | 04:24 |
| anotherjesse | when you authenticate against nova, nova asks keystone to validate the token and provide the user/tenant/roles | 04:24 |
| anotherjesse | this happens when it hits this in the middleware: https://github.com/openstack/keystone/blob/master/keystone/middleware/auth_token.py | 04:25 |
| anotherjesse | resources in swift/nova/... are marked with the user_id / tenant_id (in the db it is still called project_id) | 04:25 |
| anotherjesse | but there is no active user/project table (those are there for deprecated auth which we hope to remove in essex) | 04:26 |
| andrewbogott | So in nova, the instance table has a fields for tenant and user, but those fields relate to keystone's data -- correct? | 04:27 |
| anotherjesse | andrewbogott: correct | 04:27 |
| andrewbogott | And keystone is mandatory for essex and later, right? So I can just presume that any nova install has an associated keystone? | 04:28 |
| anotherjesse | the only change DATABASE wise - is that all nova knows now is the "foreign keys" (the user/tenant id's provided by auth_token middleware) | 04:28 |
| *** rkukura has joined #openstack-dev | 04:28 | |
| anotherjesse | and the user/project table are in keystone | 04:28 |
| anotherjesse | nova *could* keep a local cache of names/emails/... but the idea is that you probably want to backend keystone with your existing LDAP/AD system | 04:29 |
| * andrewbogott nods | 04:29 | |
| anotherjesse | so nova would always have an old cache | 04:29 |
| med_ | I'm thinking the last 20 minutes of scrollback would make a great presentation... paper... etc. | 04:29 |
| andrewbogott | Should the nova db api have a bunch of 'warning: deprecated' log messages in all those function calls? | 04:29 |
| andrewbogott | (not that I'm volunteering...) | 04:29 |
| anotherjesse | andrewbogott: ++ let's fall | 04:29 |
| anotherjesse | err | 04:29 |
| *** zaitcev has quit IRC | 04:30 | |
| anotherjesse | can you file a bug ? I'll get it fixed ;) | 04:30 |
| andrewbogott | sure. | 04:30 |
| anotherjesse | I have to run - be back in ~15 minutes to answer more keystone/nova user questions | 04:31 |
| andrewbogott | It's nearly bedtime for me anyway, so I probably have enough info for now. Thanks for your explanation. | 04:31 |
| anotherjesse | np | 04:32 |
| *** yamahata has quit IRC | 04:33 | |
| *** yamahata__ has quit IRC | 04:33 | |
| anotherjesse | med_: what would be most effective for you - presentation, video overview, we have docs like http://docs.openstack.org/trunk/openstack-identity/admin/content/what-is.html | 04:33 |
| anotherjesse | (in process of being updated from diablo to essex) | 04:33 |
| med_ | that link is good. thanks. | 04:34 |
| andrewbogott | anotherjesse: bug 949652 | 04:39 |
| uvirtbot | Launchpad bug 949652 in nova "Mark deprecated db calls as 'deprecated'" [Undecided,New] https://launchpad.net/bugs/949652 | 04:39 |
| andrewbogott | heh | 04:39 |
| *** sniperd has quit IRC | 04:40 | |
| *** sniperd has joined #openstack-dev | 04:45 | |
| *** x86brandon has quit IRC | 04:56 | |
| rmk | Traced the wsproxy problem I was having -- turned out to be an issue with setting queues durable. | 04:58 |
| rmk | Also wsproxy wasn't logging a failure to connect to the queue | 04:59 |
| rmk | Maybe current versions do | 04:59 |
| *** gkotton has joined #openstack-dev | 05:07 | |
| *** mszilagyi has quit IRC | 05:07 | |
| *** andrewsmedina has quit IRC | 05:07 | |
| *** adjohn has joined #openstack-dev | 05:09 | |
| *** sniperd has quit IRC | 05:09 | |
| *** sniperd has joined #openstack-dev | 05:17 | |
| notmyname | mnewby: thanks for the docs | 05:18 |
| mnewby | notmyname: np. Does that work, or should the saio be updated too? | 05:19 |
| mnewby | notmyname: And what would you like to see the default tox builds be? | 05:19 |
| notmyname | I think it works. we'll see how pickly gholt is :-) | 05:19 |
| mnewby | ok. sounds good :) | 05:19 |
| notmyname | mnewby: we target latest LTS (which is py26) and don't want to gate on pep8 | 05:20 |
| notmyname | I'm guessing that precise (next LTS) will have py27, but we don't/havent' targeted it yet | 05:21 |
| mnewby | notmyname: I understand. When I say 'default tox builds' I mean what developers will get when the execute 'tox' without arguments. | 05:21 |
| mnewby | the => they | 05:21 |
| notmyname | I'd say py26 and pep8 | 05:21 |
| mnewby | Ok, I'll post an updated change. | 05:21 |
| notmyname | since that's pretty similar to what ./.unitests does | 05:21 |
| mnewby | Oh, wait. That's already the default. Nevermind. | 05:21 |
| mnewby | yup | 05:22 |
| mnewby | Ok, so provided gholt is happy with the docs the change is good. | 05:22 |
| notmyname | I don't think the pep8 ignore line is the best though | 05:22 |
| notmyname | ya, I think so | 05:22 |
| mnewby | The ignore was taken from nova. What would you prefer? | 05:23 |
| notmyname | because of a bug in the pep8 tool, we don't actually pass pep8 unless you suppress that warning. but it's a valid warning in other cases | 05:23 |
| notmyname | ideally, I'd want it with no ignored warnings | 05:24 |
| notmyname | but that's a very minor thing to me | 05:24 |
| mnewby | So long as tox.ini is consistent across all projects, once the pep8 bug is fixed all projects can be updated at once. | 05:24 |
| mnewby | Is pep8 not under active development? That would seem a pretty trivial thing to fix. | 05:25 |
| notmyname | I'm fine with keeping it consistent. we'll probably still run it by hand to check submitted patches. I don't think I've ever seen anyone submit a patch that would give a false positive with the ignore yet | 05:26 |
| mnewby | It seems like a pretty rare case. I'd be hard-pressed not to spot that in a change with just my eyeballs. | 05:27 |
| notmyname | IOW, I think the patch is good for now. :-0 | 05:27 |
| notmyname | *:-) | 05:27 |
| mnewby | heh | 05:27 |
| *** relateable has joined #openstack-dev | 05:28 | |
| *** yamahata__ has joined #openstack-dev | 05:30 | |
| *** yamahata_ has joined #openstack-dev | 05:30 | |
| *** yamahata_ has left #openstack-dev | 05:32 | |
| *** hub_cap has quit IRC | 05:36 | |
| *** yamahata__ has quit IRC | 05:42 | |
| *** yamahata has joined #openstack-dev | 05:42 | |
| *** vizsla has quit IRC | 05:43 | |
| *** Asim has joined #openstack-dev | 05:45 | |
| cloudfly | anyone have any example code of an authentication query against keystone somewhere? | 05:46 |
| mnewby | Does anybody want to talk to me about DeprecationWarning? | 05:46 |
| *** dayou has joined #openstack-dev | 05:47 | |
| cloudfly | ah found it | 05:48 |
| cloudfly | danke anotherjesse | 05:48 |
| *** lloydde has joined #openstack-dev | 05:51 | |
| *** sniperd has quit IRC | 05:54 | |
| *** relateable_ has joined #openstack-dev | 05:58 | |
| *** mszilagyi has joined #openstack-dev | 05:59 | |
| *** danwent has joined #openstack-dev | 06:00 | |
| *** martine has quit IRC | 06:02 | |
| *** cp16net has quit IRC | 06:03 | |
| *** markmc has joined #openstack-dev | 06:13 | |
| *** sniperd has joined #openstack-dev | 06:18 | |
| *** zigo has joined #openstack-dev | 06:23 | |
| *** hattwick has quit IRC | 06:31 | |
| *** bepernoot has joined #openstack-dev | 06:35 | |
| *** bencherian has joined #openstack-dev | 06:38 | |
| *** relateable_ has quit IRC | 06:38 | |
| *** gkotton has quit IRC | 06:41 | |
| *** Asim has quit IRC | 06:48 | |
| *** lloydde has quit IRC | 06:48 | |
| *** littleidea has quit IRC | 06:53 | |
| *** mnewby has quit IRC | 06:54 | |
| *** sniperd has quit IRC | 06:54 | |
| *** sniperd has joined #openstack-dev | 06:55 | |
| *** zigo has quit IRC | 06:56 | |
| *** Mkenneth has joined #openstack-dev | 06:57 | |
| *** deshantm has joined #openstack-dev | 07:01 | |
| *** bepernoot has quit IRC | 07:02 | |
| *** Mkenneth1 has joined #openstack-dev | 07:04 | |
| *** Mkenneth has quit IRC | 07:06 | |
| *** dtroyer has quit IRC | 07:07 | |
| *** thingee has quit IRC | 07:10 | |
| *** zykes- has quit IRC | 07:19 | |
| *** zykes- has joined #openstack-dev | 07:21 | |
| *** deshantm has quit IRC | 07:26 | |
| *** zigo has joined #openstack-dev | 07:40 | |
| *** danwent has quit IRC | 07:41 | |
| *** mszilagyi has quit IRC | 08:02 | |
| *** bepernoot has joined #openstack-dev | 08:03 | |
| *** dneary has joined #openstack-dev | 08:09 | |
| *** dneary has quit IRC | 08:09 | |
| *** dneary has joined #openstack-dev | 08:09 | |
| *** Mkenneth1 has quit IRC | 08:10 | |
| *** mnewby has joined #openstack-dev | 08:10 | |
| *** pixelbeat has joined #openstack-dev | 08:11 | |
| *** reidrac has joined #openstack-dev | 08:20 | |
| *** Mkenneth has joined #openstack-dev | 08:22 | |
| *** sandywalsh has quit IRC | 08:24 | |
| *** bencherian has quit IRC | 08:26 | |
| *** gkotton has joined #openstack-dev | 08:37 | |
| LarsErikP | hi. Is there any document with upcoming features in Essex dashboard? | 08:37 |
| *** shevek_ has joined #openstack-dev | 08:37 | |
| *** Ryan_Lane has joined #openstack-dev | 08:38 | |
| *** sandywalsh has joined #openstack-dev | 08:39 | |
| *** Remco_ has joined #openstack-dev | 08:53 | |
| *** hattwick has joined #openstack-dev | 09:00 | |
| *** derekh has joined #openstack-dev | 09:01 | |
| *** mszilagyi has joined #openstack-dev | 09:03 | |
| *** hashar has joined #openstack-dev | 09:05 | |
| *** mnewby has quit IRC | 09:06 | |
| *** pixelbeat has quit IRC | 09:08 | |
| *** zigo has quit IRC | 09:10 | |
| *** pixelbeat has joined #openstack-dev | 09:24 | |
| *** oneiroi has joined #openstack-dev | 09:27 | |
| *** Ryan_Lane has quit IRC | 09:29 | |
| *** dayou has quit IRC | 09:35 | |
| *** dayou has joined #openstack-dev | 09:47 | |
| *** GheRivero has quit IRC | 09:56 | |
| *** hashar has left #openstack-dev | 10:00 | |
| *** adjohn has quit IRC | 10:00 | |
| *** darraghb has joined #openstack-dev | 10:01 | |
| *** maploin has joined #openstack-dev | 10:08 | |
| *** maploin has quit IRC | 10:09 | |
| *** maploin has joined #openstack-dev | 10:09 | |
| *** dayou has quit IRC | 10:13 | |
| *** gkotton has quit IRC | 10:34 | |
| *** gkotton has joined #openstack-dev | 10:34 | |
| *** gkotton has quit IRC | 10:37 | |
| *** gkotton has joined #openstack-dev | 10:41 | |
| *** apevec has joined #openstack-dev | 11:03 | |
| *** lts has joined #openstack-dev | 11:11 | |
| *** Remco_ has quit IRC | 11:14 | |
| *** Remco_ has joined #openstack-dev | 11:15 | |
| *** Remco__ has joined #openstack-dev | 11:17 | |
| *** Remco_ has quit IRC | 11:19 | |
| *** sandywalsh has quit IRC | 11:28 | |
| *** sandywalsh has joined #openstack-dev | 11:40 | |
| *** kbringard has joined #openstack-dev | 11:41 | |
| *** oneiroi has quit IRC | 11:43 | |
| *** fesnel has quit IRC | 12:07 | |
| *** cdub has quit IRC | 12:08 | |
| *** dneary has quit IRC | 12:08 | |
| *** cdub has joined #openstack-dev | 12:08 | |
| *** fesnel has joined #openstack-dev | 12:09 | |
| *** bsza has joined #openstack-dev | 12:09 | |
| *** armaan has joined #openstack-dev | 12:12 | |
| *** zykes- has quit IRC | 12:13 | |
| *** zykes- has joined #openstack-dev | 12:15 | |
| *** dneary has joined #openstack-dev | 12:18 | |
| *** dneary has quit IRC | 12:18 | |
| *** dneary has joined #openstack-dev | 12:18 | |
| *** zykes- has quit IRC | 12:25 | |
| *** zykes- has joined #openstack-dev | 12:27 | |
| *** vincentricci has joined #openstack-dev | 12:37 | |
| *** vizsla has joined #openstack-dev | 12:39 | |
| *** ncode has joined #openstack-dev | 12:42 | |
| *** ncode has joined #openstack-dev | 12:42 | |
| *** ncode has quit IRC | 12:49 | |
| *** markvoelker has joined #openstack-dev | 12:53 | |
| *** armaan has left #openstack-dev | 12:57 | |
| *** vizsla has quit IRC | 13:12 | |
| *** mdomsch has quit IRC | 13:20 | |
| *** dprince has joined #openstack-dev | 13:21 | |
| *** hashar has joined #openstack-dev | 13:22 | |
| *** Mkenneth has quit IRC | 13:23 | |
| *** sniperd has quit IRC | 13:24 | |
| *** andrewsmedina has joined #openstack-dev | 13:26 | |
| *** vincentricci has joined #openstack-dev | 13:33 | |
| *** vincentricci has left #openstack-dev | 13:33 | |
| *** dprince has quit IRC | 13:40 | |
| *** dprince has joined #openstack-dev | 13:40 | |
| *** andrewsmedina has quit IRC | 13:41 | |
| *** andrewsmedina has joined #openstack-dev | 13:41 | |
| *** PotHix has joined #openstack-dev | 13:50 | |
| *** CatKiller has quit IRC | 13:53 | |
| *** martine has joined #openstack-dev | 13:57 | |
| *** lts has quit IRC | 14:04 | |
| *** CatKiller has joined #openstack-dev | 14:08 | |
| *** littleidea has joined #openstack-dev | 14:14 | |
| *** sniperd has joined #openstack-dev | 14:19 | |
| *** rnirmal has joined #openstack-dev | 14:22 | |
| *** Remco__ has quit IRC | 14:22 | |
| *** dneary has quit IRC | 14:23 | |
| *** Remco_ has joined #openstack-dev | 14:23 | |
| *** Glacee has joined #openstack-dev | 14:25 | |
| *** dtroyer has joined #openstack-dev | 14:26 | |
| *** maploin has quit IRC | 14:26 | |
| *** maploin has joined #openstack-dev | 14:27 | |
| *** maploin has quit IRC | 14:27 | |
| *** maploin has joined #openstack-dev | 14:27 | |
| *** Remco_ has quit IRC | 14:28 | |
| *** Mandell_ has quit IRC | 14:35 | |
| *** aweiss1 has joined #openstack-dev | 14:35 | |
| *** aweiss1 has quit IRC | 14:39 | |
| *** aweiss has joined #openstack-dev | 14:40 | |
| *** ccorrigan has joined #openstack-dev | 14:40 | |
| *** eglynn_ has quit IRC | 14:42 | |
| *** davlap has joined #openstack-dev | 14:43 | |
| *** ncode has joined #openstack-dev | 14:45 | |
| *** ncode has joined #openstack-dev | 14:45 | |
| *** eglynn has joined #openstack-dev | 14:50 | |
| *** Mandell has joined #openstack-dev | 14:55 | |
| *** deshantm has joined #openstack-dev | 14:57 | |
| *** jdg has joined #openstack-dev | 15:02 | |
| *** zaitcev has joined #openstack-dev | 15:07 | |
| *** lloydde has joined #openstack-dev | 15:08 | |
| *** lloydde has quit IRC | 15:08 | |
| *** rods has joined #openstack-dev | 15:09 | |
| *** ayoung has joined #openstack-dev | 15:10 | |
| *** aweiss has quit IRC | 15:13 | |
| *** AlanClark has joined #openstack-dev | 15:20 | |
| *** dneary has joined #openstack-dev | 15:22 | |
| *** dneary has quit IRC | 15:22 | |
| *** dneary has joined #openstack-dev | 15:22 | |
| *** ywu-os has joined #openstack-dev | 15:23 | |
| *** lloydde has joined #openstack-dev | 15:23 | |
| *** bsza has quit IRC | 15:28 | |
| *** jakedahn has quit IRC | 15:29 | |
| *** jakedahn_zz is now known as jakedahn | 15:29 | |
| *** mancdaz has quit IRC | 15:29 | |
| *** ayoung has quit IRC | 15:39 | |
| *** deshantm has quit IRC | 15:39 | |
| *** glenc has joined #openstack-dev | 15:41 | |
| *** eglynn has quit IRC | 15:41 | |
| *** GheRivero_ has joined #openstack-dev | 15:42 | |
| *** glenc_ has quit IRC | 15:42 | |
| *** bsza has joined #openstack-dev | 15:43 | |
| *** mancdaz has joined #openstack-dev | 15:45 | |
| *** jdg has quit IRC | 15:46 | |
| *** ayoung has joined #openstack-dev | 15:50 | |
| *** lloydde has quit IRC | 15:50 | |
| *** dolphm has joined #openstack-dev | 15:53 | |
| *** eglynn has joined #openstack-dev | 15:54 | |
| *** mattray has joined #openstack-dev | 15:58 | |
| *** aweiss has joined #openstack-dev | 15:58 | |
| *** zzed has joined #openstack-dev | 15:59 | |
| *** dtroyer has quit IRC | 16:05 | |
| *** reidrac has quit IRC | 16:06 | |
| *** heckj has joined #openstack-dev | 16:07 | |
| *** bencherian has joined #openstack-dev | 16:08 | |
| *** anotherjesse is now known as anotherjesse_zz | 16:12 | |
| *** hhoover has joined #openstack-dev | 16:14 | |
| *** danwent has joined #openstack-dev | 16:15 | |
| *** hub_cap has joined #openstack-dev | 16:19 | |
| *** dtroyer has joined #openstack-dev | 16:19 | |
| *** dwalleck has joined #openstack-dev | 16:20 | |
| *** dtroyer has quit IRC | 16:25 | |
| *** dtroyer has joined #openstack-dev | 16:25 | |
| *** sniperd has quit IRC | 16:28 | |
| *** bepernoot has quit IRC | 16:30 | |
| *** JStoker_ has quit IRC | 16:30 | |
| *** maplebed has joined #openstack-dev | 16:33 | |
| *** maplebed has quit IRC | 16:33 | |
| *** maplebed has joined #openstack-dev | 16:33 | |
| *** bencherian has quit IRC | 16:37 | |
| *** JStoker_ has joined #openstack-dev | 16:41 | |
| *** jaypipes has quit IRC | 16:44 | |
| *** dwalleck_ has joined #openstack-dev | 16:45 | |
| *** danwent has quit IRC | 16:46 | |
| *** dwalleck has quit IRC | 16:49 | |
| *** danwent has joined #openstack-dev | 16:53 | |
| *** dwalleck_ has quit IRC | 16:53 | |
| *** anotherjesse_zz is now known as anotherjesse | 16:54 | |
| *** sniperd has joined #openstack-dev | 16:56 | |
| *** lloydde has joined #openstack-dev | 17:00 | |
| *** rnirmal has quit IRC | 17:00 | |
| *** rnirmal has joined #openstack-dev | 17:00 | |
| *** dwalleck has joined #openstack-dev | 17:02 | |
| *** roge has joined #openstack-dev | 17:02 | |
| *** mdomsch has joined #openstack-dev | 17:03 | |
| *** dwalleck has quit IRC | 17:05 | |
| *** infernix has joined #openstack-dev | 17:05 | |
| *** cp16net has joined #openstack-dev | 17:06 | |
| anotherjesse | comstud: around? got opinions on if/where I should move fake memcache - https://review.openstack.org/#change,5042 makes use of it not just for testing | 17:06 |
| *** mattray has quit IRC | 17:07 | |
| *** mattray has joined #openstack-dev | 17:07 | |
| *** mattray has quit IRC | 17:07 | |
| *** mattray has joined #openstack-dev | 17:07 | |
| *** sniperd has quit IRC | 17:07 | |
| *** dolphm has quit IRC | 17:09 | |
| *** sniperd has joined #openstack-dev | 17:16 | |
| annegentle | anotherjesse: did anything in the index.html change surrounding the Google search? See this bug: | 17:18 |
| annegentle | https://bugs.launchpad.net/openstack-manuals/+bug/950092 | 17:18 |
| uvirtbot | Launchpad bug 950092 in openstack-manuals "Search results on docs.openstack.org now display Google Ads and shouldn't" [High,Confirmed] | 17:18 |
| annegentle | anotherjesse: if anyone can take a look, I'd appreciate it! | 17:18 |
| *** dwalleck has joined #openstack-dev | 17:19 | |
| anotherjesse | annegentle: looking - sending a private message | 17:19 |
| comstud | anotherjesse: i am around.. looking | 17:20 |
| anotherjesse | comstud: vishy is around too - we can yell at him for thoughts on it | 17:21 |
| comstud | ah yes, this one | 17:21 |
| anotherjesse | annegentle: if you could answer private message about details | 17:21 |
| *** dolphm has joined #openstack-dev | 17:24 | |
| *** spiffxp has joined #openstack-dev | 17:25 | |
| anotherjesse | annegentle: updated bug with details - don't think it is something that is controlled from code - but from google | 17:28 |
| comstud | anotherjesse: left a comment | 17:31 |
| *** troytoman-away is now known as troytoman | 17:32 | |
| *** jdurgin has joined #openstack-dev | 17:33 | |
| *** dolphm has quit IRC | 17:35 | |
| anotherjesse | vishy: http://evan.prodromou.name/Software/Python/LRUCache | 17:37 |
| anotherjesse | for later | 17:37 |
| *** gkotton has quit IRC | 17:37 | |
| comstud | i have a bone to pick with vishy, also. one of his commits made me waste a lot of time last night :) | 17:40 |
| anotherjesse | just one? | 17:41 |
| comstud | haha | 17:41 |
| comstud | just one last night | 17:41 |
| comstud | he caused a messaging change by changing RequestContext | 17:41 |
| anotherjesse | ah, hence the kwargs patch | 17:42 |
| comstud | yep | 17:42 |
| comstud | it wasn't his fault it was so difficult to track down | 17:42 |
| anotherjesse | sure it is | 17:43 |
| comstud | ok | 17:43 |
| comstud | yes it is | 17:43 |
| comstud | or waldon. | 17:43 |
| *** maploin has quit IRC | 17:43 | |
| comstud | it's waldon's fault. | 17:43 |
| *** anotherjesse is now known as anotherjesse_zz | 17:45 | |
| *** dneary has quit IRC | 17:48 | |
| *** aweiss has quit IRC | 17:48 | |
| *** vincentricci has joined #openstack-dev | 17:54 | |
| *** davlap has quit IRC | 17:56 | |
| *** bepernoot has joined #openstack-dev | 17:57 | |
| *** dtroyer_ has joined #openstack-dev | 17:57 | |
| *** gyee has joined #openstack-dev | 17:58 | |
| *** bepernoot has quit IRC | 17:59 | |
| *** mdomsch has quit IRC | 17:59 | |
| *** johnpostlethwait has joined #openstack-dev | 18:00 | |
| *** johnpostlethwait has quit IRC | 18:00 | |
| *** johnpostlethwait has joined #openstack-dev | 18:01 | |
| *** dtroyer has quit IRC | 18:01 | |
| *** dtroyer_ is now known as dtroyer | 18:01 | |
| *** pixelbeat has quit IRC | 18:04 | |
| *** dwalleck has quit IRC | 18:07 | |
| vishy | comstud: uh oh | 18:07 |
| vishy | comstud: which one? | 18:07 |
| *** gkotton has joined #openstack-dev | 18:09 | |
| comstud | the context change | 18:09 |
| comstud | switching the determination of auth strategy | 18:09 |
| comstud | you removed 'strategy' from RequestContext | 18:09 |
| comstud | since it disappeared as a keyword arg in __init__() | 18:09 |
| *** dalang has joined #openstack-dev | 18:10 | |
| comstud | rpc code calling to_dict() barfed on an old message that still had 'strategy' in it | 18:10 |
| comstud | vishy: ^ | 18:10 |
| comstud | unfortunately, no exception was logged | 18:10 |
| comstud | and rpc code stopped consuming messages | 18:11 |
| comstud | so it'd pull a message off the queue and just silently halt. | 18:11 |
| vishy | ouch | 18:11 |
| comstud | i submitted a 2nd fix for that | 18:11 |
| vishy | there was an old message? | 18:11 |
| comstud | yep | 18:12 |
| comstud | in our queue | 18:12 |
| vishy | comstud: note to self: don't mess with req.context | 18:12 |
| comstud | i guess we restarted as just the right time | 18:12 |
| comstud | as/at | 18:12 |
| comstud | vishy:https://review.openstack.org/#change,5072 | 18:12 |
| comstud | that should address it | 18:12 |
| comstud | at least.. this specific case :) | 18:12 |
| comstud | that's what I monkey patched in to get things flowing again | 18:13 |
| *** sniperd has quit IRC | 18:15 | |
| *** thingee has joined #openstack-dev | 18:15 | |
| *** jakedahn is now known as jakedahn_zz | 18:16 | |
| *** sdake has quit IRC | 18:16 | |
| *** sdake has joined #openstack-dev | 18:16 | |
| *** johnpostlethwait has quit IRC | 18:18 | |
| *** dwalleck has joined #openstack-dev | 18:19 | |
| *** johnpostlethwait has joined #openstack-dev | 18:19 | |
| *** dwalleck has quit IRC | 18:19 | |
| *** bencherian has joined #openstack-dev | 18:20 | |
| *** Ryan_Lane has joined #openstack-dev | 18:21 | |
| *** shevek_ has quit IRC | 18:22 | |
| *** mdomsch has joined #openstack-dev | 18:23 | |
| *** johnpostlethwait has quit IRC | 18:26 | |
| *** johnpostlethwait has joined #openstack-dev | 18:27 | |
| *** johnpostlethwait has quit IRC | 18:29 | |
| *** johnpostlethwait has joined #openstack-dev | 18:29 | |
| *** rnirmal has quit IRC | 18:33 | |
| *** johnpostlethwait has quit IRC | 18:34 | |
| *** johnpostlethwait has joined #openstack-dev | 18:35 | |
| *** johnpostlethwait has quit IRC | 18:37 | |
| *** johnpostlethwait has joined #openstack-dev | 18:38 | |
| *** novas0x2a|laptop has joined #openstack-dev | 18:39 | |
| *** lloydde has quit IRC | 18:40 | |
| *** rnirmal has joined #openstack-dev | 18:47 | |
| *** rnirmal has quit IRC | 18:48 | |
| sleepsonzzz | heckj - what is the recommended way for configuring a service catalog right now? is default_catalog.templates no longer required? | 18:48 |
| *** mattstep has quit IRC | 18:53 | |
| *** mattstep has joined #openstack-dev | 18:53 | |
| sleepsonzzz | heckl: should we change the way devstack is configuring the endpoints? | 18:55 |
| *** mattray has quit IRC | 19:02 | |
| *** dwalleck has joined #openstack-dev | 19:03 | |
| *** sandywalsh has quit IRC | 19:03 | |
| *** dwalleck has quit IRC | 19:04 | |
| *** dwalleck has joined #openstack-dev | 19:04 | |
| *** dwalleck has quit IRC | 19:06 | |
| *** vizsla has joined #openstack-dev | 19:08 | |
| *** darraghb has quit IRC | 19:12 | |
| *** vizsla has quit IRC | 19:13 | |
| *** mattray has joined #openstack-dev | 19:19 | |
| *** mnewby has joined #openstack-dev | 19:23 | |
| *** griff has joined #openstack-dev | 19:25 | |
| *** hub_cap has quit IRC | 19:25 | |
| *** hub_cap has joined #openstack-dev | 19:26 | |
| heckj | sleepsonzzz: still using service catalog right now - we just slipped the SQL backend in for catalog prior to E4, but there's some larger "what are we actually modeling?" issues around the existing API. | 19:30 |
| tr3buchet | vishy: i see | 19:30 |
| vishy | tr3buchet: so the deal is that release actually forces a dhcp release which adds the ip back into the pool | 19:30 |
| tr3buchet | vishy: whereas update just stops dnsmasq from responding? | 19:31 |
| vishy | if we are tearing down a network, we don't want to give the ip back | 19:31 |
| vishy | tr3buchet: update just updates the list of ips that dnsmasq is responding to | 19:31 |
| *** mnewby has quit IRC | 19:31 | |
| tr3buchet | ok, makes sense | 19:31 |
| *** mnewby has joined #openstack-dev | 19:31 | |
| tr3buchet | that function is nasty though... | 19:31 |
| vishy | tr3buchet: yeah :( | 19:32 |
| sleepsonzzz | heckj ah ok. was looking at https://bugs.launchpad.net/bugs/938243 - so for now, deployers will still need to use keystone service-create as well as the default_catalog.templates? | 19:32 |
| uvirtbot | Launchpad bug 938243 in keystone "keystone service-create seems redundant with default_catalog.templates" [Medium,Invalid] | 19:32 |
| tr3buchet | scaling at 2n is bad | 19:32 |
| vishy | tr3buchet: I will be happy when melange takes that over | 19:32 |
| tr3buchet | vishy: me too | 19:32 |
| vishy | tr3buchet: that said, it is pretty rare for a vm to have more than one ip on the same network | 19:32 |
| vishy | tr3buchet: so it will almost never get called twice | 19:32 |
| tr3buchet | vishy: it's more the size of the network that bothers me | 19:33 |
| tr3buchet | calling it once when there are 250 IPs in th enetwork results in 501 db calls | 19:33 |
| mnewby | mtaylor: Can you help with landing the tox patch for swift? | 19:34 |
| *** sandywalsh has joined #openstack-dev | 19:34 | |
| vishy | tr3buchet: ah actually there is an error | 19:36 |
| vishy | tr3buchet: I'm just noticing that deallocate_fixed_ip is using teardown | 19:36 |
| heckj | sleepsonzzz: for now, expecting deployers to use templates, but we need to have some clarity in a (currently missing) catalog.core.Driver class as to what it needs to do and how it's allowed to responsd (internal to keystone python API). i.e. specifically returning NotImplemented or NotAllowedToDoThatWithThisBackend errors. | 19:36 |
| tr3buchet | vishy: that's how it was before | 19:36 |
| vishy | tr3buchet: but i clobberred the dhcp release functionality | 19:37 |
| tr3buchet | vishy: ah i see | 19:37 |
| vishy | tr3buchet: so deallocate_fixed_ip, should it be tearing down network every time? | 19:39 |
| tr3buchet | well | 19:39 |
| vishy | tr3buchet: seems like it should only happen if it is the last fixed ip in the network? | 19:39 |
| tr3buchet | the way it was written, i separated out setup and teardown from allocate and deallocate ip | 19:39 |
| tr3buchet | so i took was deallocate was doing and put it into teardown | 19:40 |
| tr3buchet | some of what it was doing | 19:40 |
| tr3buchet | so that it could be called by setup_networks | 19:40 |
| vishy | tr3buchet: so the issue is, update_dhcp doesn't need to be called there | 19:41 |
| vishy | because it is called by release ip | 19:41 |
| vishy | tr3buchet: so i actually think that it should be called in release_ip instead if it is necessary to call it | 19:41 |
| sleepsonzzz | heckj - ok, makes sense, though that bug should probably remain active until such a time that users can actually have a usable non-redundant configuration method. Unless I'm mis-understanding, and using the template backend now obviates the need to use keystone service-create. | 19:42 |
| tr3buchet | hmmm | 19:43 |
| tr3buchet | vishy: do you know what deallocate fixed ip was called dhcp release? | 19:43 |
| tr3buchet | s/what/why/ | 19:43 |
| vishy | tr3buchet: it didn't | 19:43 |
| vishy | it calls it if force_dhcp_release is enabled | 19:44 |
| tr3buchet | right | 19:44 |
| tr3buchet | it still does exactly that | 19:44 |
| vishy | well it does before my change | 19:44 |
| tr3buchet | on now "that" is called _teardown_network_on_host | 19:44 |
| vishy | but the deal is teardown needs to update_dhcp | 19:44 |
| vishy | not force_release | 19:44 |
| tr3buchet | it seems like to me we can revert deallocate to call dhcp release as it did before | 19:45 |
| *** dolphm has joined #openstack-dev | 19:45 | |
| heckj | sleepsonzzz: the template backend for services does mean that the service commands aren't needed - which is why I marked it invalid. The commands are still needed for the SQL backend, but the equiv of "service create" for the SQL backend is just adding lines into the template config file. | 19:45 |
| *** reed_ has joined #openstack-dev | 19:46 | |
| tr3buchet | vishy: and then setup_networks_on_host doesn't have to loop through all of the IPs | 19:47 |
| tr3buchet | vishy: make sense? | 19:47 |
| sleepsonzzz | heckj - ok sweet, so we should remove service-create calls from devstack? | 19:47 |
| vishy | tr3buchet: hold on about to put up a patch | 19:47 |
| heckj | sleepsonzzz: yeah - or make them variable on having the SQL backend for the catalog | 19:48 |
| *** hub_cap has quit IRC | 19:50 | |
| *** hub_cap has joined #openstack-dev | 19:50 | |
| sleepsonzzz | heckj - cool, I'll take a look - thanks for the help | 19:51 |
| adam_g | heckj: so the default template config needs lines added for the services that are created in keystone_data.sh? IIRC, service-create is needed to define the services, but they are lost on keystone restart when using kvs backend. can those also come the template backend? | 19:54 |
| *** jakedahn_zz is now known as jakedahn | 19:54 | |
| heckj | adam_g: I'd need to re-read the code to verify, but I thought that with the catalog back end, there wasn't any KVS store for the services themselves - they were simply rendered directly from the template and that was that. | 19:55 |
| vishy | tr3buchet: I'm starting to feel like passing vif and address into teardown_network is superfluous | 19:56 |
| adam_g | heckj: ya, i'd have to confirm but i *think* the services is non-persitent. in any case, i think the backend should at least be configurable in devstack to document the different flavors | 20:02 |
| tr3buchet | i agree | 20:02 |
| tr3buchet | vishy: ^ | 20:02 |
| vishy | tr3buchet: ok one thing I don't understand, why is flat manager calling super for allocate and deallocate? | 20:02 |
| *** jakedahn is now known as jakedahn_zz | 20:02 | |
| vishy | tr3buchet: ah so that force release thing needs to be moved | 20:02 |
| tr3buchet | vishy: you got it? | 20:03 |
| vishy | tr3buchet: almost | 20:03 |
| vishy | tr3buchet: so the force dhcp release thing being in the superclass is kind of nasty | 20:03 |
| vishy | but i won't move it right now | 20:03 |
| vishy | tr3buchet: the problem is flat should call teardown in deallocate | 20:04 |
| vishy | whereas the dhcp managers should call it in release | 20:04 |
| vishy | tr3buchet: i guess flatdhcp could call it in both places, but that seems a litle nasty | 20:05 |
| tr3buchet | vishy: yeah | 20:06 |
| tr3buchet | vishy: it's kindof hard to grow things when there are 3 | 20:06 |
| tr3buchet | well 4 places | 20:07 |
| tr3buchet | that the code can go | 20:07 |
| tr3buchet | people kept updating stuff in one and not the others | 20:07 |
| tr3buchet | so i put as much stuff in the superclass as possible | 20:07 |
| tr3buchet | the flag ensures that bit wasn't called in flat | 20:07 |
| vishy | tr3buchet: so I'm trying to think if calling it in deallocate will cause problems one sec | 20:10 |
| tr3buchet | kk | 20:10 |
| *** zzed has quit IRC | 20:13 | |
| *** adjohn has joined #openstack-dev | 20:13 | |
| cloudfly | hrmm so i am trying to build a pam auth module to auth against keystone.. the question is... is that really the right mechanism? Can keystone be made to do gssapi or some other automated token passing? for seamless authentication using the API token? is that even a good idea? | 20:15 |
| vishy | tr3buchet: ok i think i have a solution | 20:16 |
| *** pixelbeat has joined #openstack-dev | 20:17 | |
| tr3buchet | yay | 20:17 |
| vishy | tr3buchet: which is to only allow allocated fixed_ips in dnsmasq | 20:17 |
| vishy | tr3buchet: so it will stop putting them in the conf file immediately, but it won't disassociate them in the db, then I don't have to call teardown_network in release | 20:18 |
| *** bencherian has quit IRC | 20:18 | |
| *** aweiss has joined #openstack-dev | 20:19 | |
| adam_g | vishy: --auth_strategy is now required to be defined on both nova-api and nova-compute nodes, correct? | 20:20 |
| tr3buchet | vishy: wouldn't it be weird for them to be associated but not in the conf file? | 20:20 |
| *** adjohn has quit IRC | 20:20 | |
| vishy | adam_g: i would define it everywhere, yes | 20:20 |
| vishy | tr3buchet: well if it is no longer allocated it means it just hasn't been marked released yet by dnsmasq | 20:21 |
| vishy | tr3buchet: so I think it is fine for it to not be in conf anymore. It shouldn't be handing out new leases for that ip anyway | 20:21 |
| *** adjohn has joined #openstack-dev | 20:21 | |
| *** sniperd has joined #openstack-dev | 20:21 | |
| tr3buchet | vishy: right, not until it gets allocated again | 20:22 |
| *** sniperd has quit IRC | 20:22 | |
| vishy | tr3buchet: updated https://review.openstack.org/#change,5038,patchset=5 | 20:29 |
| tr3buchet | vishy: i don't know about allocated=True | 20:30 |
| tr3buchet | in network_get_associated_fixed_ips() | 20:30 |
| tr3buchet | it changes the function | 20:30 |
| tr3buchet | now it's only allocated/associated ips, instead of just associated ips | 20:31 |
| vishy | tr3buchet: correct, but it is used in exactly one place, to get a list of ips for dhcp | 20:32 |
| vishy | tr3buchet: I could add an optional param and pass it in in all three places | 20:32 |
| vishy | or i could change the name of the method | 20:33 |
| tr3buchet | if it's only used in one place i think it's fine | 20:33 |
| vishy | * one file (three places in the file) | 20:33 |
| tr3buchet | i just didnt' want to break other stuff | 20:33 |
| tr3buchet | vishy: i thought you didn't want to teardown_network_on_host in deallocate | 20:34 |
| vishy | tr3buchet: That was my initial thought, but now that we are keying dhcp leases off of allocated | 20:34 |
| vishy | tr3buchet: i don't have to call it in release_fixed_ip any more | 20:34 |
| tr3buchet | ah ok | 20:35 |
| vishy | so I think it is fine there | 20:35 |
| vishy | I'm just going to check one possible edge case scenario and make sure nova doesn't choke | 20:38 |
| vishy | that is restarting dnsmasq between deallocate and release | 20:38 |
| vishy | although the odds of that occurring in production are pretty damn low | 20:38 |
| termie | jeblair, mtaylor: this is passing for me locally using devstack, any way to get more info about what is going wrong from jenkins? https://review.openstack.org/#change,4659 | 20:41 |
| *** lloydde has joined #openstack-dev | 20:42 | |
| termie | jeblair, mtaylor: the only thing i can think of is that the policy.json file is somehow not being found | 20:42 |
| *** zzed has joined #openstack-dev | 20:43 | |
| *** zzed has quit IRC | 20:47 | |
| *** dalang has quit IRC | 20:47 | |
| *** dubsquared has joined #openstack-dev | 20:48 | |
| mtaylor | termie: looking | 20:52 |
| termie | ('cause that is the only new thing introduced in teh change) | 20:53 |
| vishy | tr3buchet: I split the change into two patches, also verified that fixes work in lab environment | 20:54 |
| tr3buchet | awesome | 20:55 |
| tr3buchet | what's the other one? | 20:55 |
| *** GheRivero_ has quit IRC | 20:57 | |
| *** dprince has quit IRC | 20:58 | |
| *** dolphm has quit IRC | 20:59 | |
| *** zzed has joined #openstack-dev | 21:01 | |
| *** dolphm has joined #openstack-dev | 21:02 | |
| andrewbogott | So... if I want to talk to keystone from my nova extension, should I import bits from python-keystoneclient? I don't see a lot of examples of keystone use at the moment. | 21:03 |
| *** zzed has quit IRC | 21:09 | |
| andrewbogott | atm I am trying to do something very simple: given a product id, I want to verify that it corresponds to an actual project. | 21:09 |
| andrewbogott | um project aka tenant | 21:09 |
| *** dubsquared has quit IRC | 21:09 | |
| *** griff has quit IRC | 21:10 | |
| tr3buchet | vishy: what is the other merge? | 21:10 |
| *** mattray has quit IRC | 21:10 | |
| mtaylor | termie: I'm guessing https://jenkins.openstack.org/job/gate-integration-tests-devstack-vm/2327/artifact/logs/syslog.txt didn't help? | 21:13 |
| mtaylor | termie: jeblair is afk - when he gets back, I think we can do a key injection for you on to that server do you can take a peek | 21:13 |
| *** jakedahn_zz is now known as jakedahn | 21:15 | |
| *** mattray has joined #openstack-dev | 21:16 | |
| *** relateable has quit IRC | 21:17 | |
| *** dubsquared has joined #openstack-dev | 21:17 | |
| *** dalang has joined #openstack-dev | 21:21 | |
| *** dwalleck has joined #openstack-dev | 21:22 | |
| *** zzed has joined #openstack-dev | 21:22 | |
| cp16net | i am getting an error right now after starting up devstack when creating an instance. 2012-03-08 13:25:27 ERROR nova.compute.manager [-] [instance: 3687c7b5-d94f-4224-a8bb-286e55d0dbd3] Instance failed network setup | 21:26 |
| cp16net | i thought the network was setup after running ./stack.sh | 21:27 |
| *** bsza1 has joined #openstack-dev | 21:27 | |
| cp16net | anyone have any pointers on how i can debug this? | 21:27 |
| *** bsza has quit IRC | 21:30 | |
| *** mattray has quit IRC | 21:31 | |
| apevec | heckj, is 3 x +1 enough to push this one? https://review.openstack.org/#change,5027 | 21:31 |
| termie | mtaylor: yeah the syslog only shows glance stuff | 21:31 |
| vishy | tr3buchet: it is listed as a dependent branch | 21:32 |
| tr3buchet | ah ok | 21:32 |
| tr3buchet | ah | 21:32 |
| tr3buchet | sure enough | 21:32 |
| termie | apevec: approved | 21:32 |
| tr3buchet | i was scrolled down | 21:32 |
| comstud | vishy: can you look @ https://review.openstack.org/#change,5069 | 21:33 |
| apevec | termie, thanks | 21:33 |
| vishy | comstud: what is importing signer? | 21:34 |
| comstud | auth.manager | 21:34 |
| comstud | it's in the description | 21:34 |
| comstud | commit reason | 21:34 |
| vishy | comstud: signer is useless without boto, so that seems a little odd. | 21:34 |
| comstud | vishy: yeah, alternative is to mod auth.manager | 21:35 |
| vishy | comstud: I think i like that better | 21:35 |
| comstud | essentially i'll have to: | 21:35 |
| *** dolphm has quit IRC | 21:35 | |
| comstud | try: | 21:35 |
| comstud | import auth.signer | 21:35 |
| comstud | tho | 21:35 |
| vishy | i mean the code is all going boom in a couple of weeks | 21:35 |
| vishy | so i guess i don't care that much | 21:35 |
| *** dwalleck_ has joined #openstack-dev | 21:35 | |
| comstud | yeah, i commented on that | 21:36 |
| comstud | that's why i took the simple approach | 21:36 |
| *** aweiss1 has joined #openstack-dev | 21:36 | |
| comstud | it affected a deploy last night | 21:36 |
| comstud | but mostly because things were importing auth.manager | 21:36 |
| comstud | that didn't need to import them | 21:36 |
| tr3buchet | vishy: are you finished with both patches now? | 21:36 |
| comstud | and I fixed those in separate patches | 21:37 |
| comstud | but the bin/nova-manage problem remains | 21:37 |
| comstud | which imports auth.manager but has no need for auth.signer | 21:37 |
| *** dolphm has joined #openstack-dev | 21:38 | |
| vishy | tr3buchet: yes | 21:38 |
| tr3buchet | k | 21:38 |
| vishy | although perhaps i should run the whole suite of unit tests :) | 21:38 |
| *** dwalleck has quit IRC | 21:39 | |
| *** aweiss has quit IRC | 21:39 | |
| comstud | jenkins will do that for you | 21:39 |
| comstud | :) | 21:39 |
| markmc | vishy, on auth.manager going boom ... there's no keystone equivalent of 'project zipfile' is there? | 21:39 |
| vishy | markmc: there is a horizon equivalent | 21:42 |
| markmc | vishy, ah, interesting - thanks | 21:42 |
| vishy | markmc: and a devstack equivalent | 21:42 |
| andrewbogott | Surely one of you has advice about calling keystone from within nova :) | 21:42 |
| markmc | vishy, heh, ok :) | 21:42 |
| soren | vishy: Sorry, what is going to go boom in a couple of weeks? | 21:43 |
| hub_cap | y2k | 21:44 |
| vishy | soren: deprecated auth! | 21:44 |
| vishy | (thank god) | 21:44 |
| soren | vishy: I thought we decided to keep that for Essex? | 21:44 |
| vishy | soren: hence the couple of weeks | 21:45 |
| vishy | soren: (read as soon as folsom opens) | 21:45 |
| soren | Ah. | 21:45 |
| vishy | markmc: did you see that we switched devstack over to using config syntax? | 21:45 |
| markmc | vishy, oh, awesome - my bad for not getting to it | 21:46 |
| vishy | markmc: also, did you see the branch that added tools/conf/generate.sh | 21:46 |
| vishy | markmc: I modified a script written by someone else. I don't know if i like the way it is doing it, but it seems like a valuable thing to have | 21:46 |
| markmc | vishy, I reviewed an early version of the generate stuff, haven't looked since | 21:47 |
| markmc | vishy, agree I didn't like it too much, but generating a default nova.conf from the cfg schemas makes total sense | 21:47 |
| vishy | markmc: can you take a look at it and see if there is anything in there? The output goes into etc/nova/nova.conf.sample | 21:47 |
| vishy | markmc: I think i properly formatted multistring and list options | 21:47 |
| markmc | vishy, latest is in master? nothing pending in gerrit? | 21:48 |
| vishy | markmc: my changes merged already | 21:48 |
| markmc | vishy, cool, looking | 21:48 |
| *** zzed has quit IRC | 21:50 | |
| hub_cap | can soomeone look @ my novaclient patch? it fixes the proxy_token behavior to work w/ the new keystone api for belongsTo https://review.openstack.org/#change,5018 | 21:50 |
| markmc | vishy, stuff like lock_path="/Users/vishvananda/cache/stack/nova/nova/../" is unfortunate | 21:51 |
| jeblair | termie: ssh jenkins@50.57.66.241 | 21:52 |
| markmc | vishy, maybe we can add a flag to cfg to disable interpolation when looking up the default | 21:52 |
| vishy | markmc: oh I didn't even notice that | 21:52 |
| jeblair | termie: your key from launchpad has been installed there | 21:52 |
| vishy | markmc: i thought it would show without interpolation | 21:52 |
| markmc | vishy, what is this alien /Users anyway? | 21:53 |
| * markmc frowns | 21:53 | |
| jeblair | termie: hopefully https://bugs.launchpad.net/devstack/+bug/945231 will improve this | 21:53 |
| uvirtbot | Launchpad bug 945231 in devstack "screen output should be logged" [High,In progress] | 21:53 |
| vishy | markmc: bsd-- | 21:53 |
| vishy | aka osx | 21:53 |
| markmc | vishy, :) | 21:53 |
| *** reed_ has quit IRC | 21:54 | |
| markmc | vishy, it looks pretty reasonable now, cool stuff | 21:55 |
| *** ayoung has quit IRC | 21:55 | |
| cp16net | i keep having a problem starting the network when attempting to grab file lock for "iptables" for method apply. | 21:55 |
| vishy | markmc: did you want to do a quick patch to remove the interpolation? | 21:56 |
| markmc | vishy, will do | 21:56 |
| vishy | cp16net: you should be able to delete stale lockfiles. I'm assuming you are using an old version of the code | 21:56 |
| vishy | cp16net: since the new code automatically does that | 21:56 |
| vishy | cp16net: you need to find out where lock_path is putting lock files | 21:56 |
| cp16net | i've just redone ./stack | 21:57 |
| cp16net | which should get all the latest code | 21:57 |
| cp16net | and i renabled RECLONE=yes in the stack.sh | 21:57 |
| vishy | cp16net: you are using devstack? | 21:57 |
| cp16net | yes | 21:57 |
| *** reed_ has joined #openstack-dev | 21:57 | |
| cp16net | vishy: i guess these are the lock files? ubuntu.Dummy-4-4142 | 21:59 |
| vishy | cp16net: oddness, it is supposed to clean them up | 21:59 |
| cp16net | it even says it is cleaning them up | 21:59 |
| vishy | cp16net: there should be ones called nova.lock in there | 21:59 |
| vishy | in /opt/stack/nova | 21:59 |
| cp16net | http://paste.openstack.org/show/7548/ | 21:59 |
| vishy | cp16net: perhaps you ran as root at some point? | 21:59 |
| termie | hmm, it seems like it wants to use the wrong policy driver... oh i bet devstack sets that | 22:00 |
| cp16net | i dont think i have run as root other than what devstack kicks off | 22:00 |
| termie | hrm, annoying recursive dependency stuff | 22:00 |
| vishy | cp16net: can you delete the file manually? | 22:00 |
| termie | (when i tested locally i did a merge on the configs but devstack would normally just overwrite) | 22:01 |
| cp16net | yeah i did delete them | 22:02 |
| *** hub_cap has quit IRC | 22:02 | |
| *** mattray has joined #openstack-dev | 22:02 | |
| cp16net | and then it crashed network | 22:02 |
| cp16net | because it was holding on to the file | 22:02 |
| cp16net | i restart network and it creates a new dummy file and stops at the same point as before | 22:03 |
| vishy | cp16net: do you have another nova process running somewhere? | 22:03 |
| vishy | cp16net: try killall screen | 22:04 |
| vishy | followed by deleting lock files | 22:04 |
| vishy | followed by ./rejoin-stack.sh | 22:04 |
| cp16net | i only had a single screen session running | 22:04 |
| cp16net | i killed it | 22:04 |
| termie | jeblair: thanks a bunch, totally found the problem | 22:05 |
| *** aweiss1 has quit IRC | 22:05 | |
| cp16net | and i am stuck in a holding pattern again | 22:06 |
| cp16net | it created 2 dummy files | 22:06 |
| cp16net | in /opt/stack/nova | 22:06 |
| jeblair | termie: groovy! | 22:07 |
| *** andrewsmedina has quit IRC | 22:10 | |
| *** hashar has quit IRC | 22:15 | |
| *** markmc has quit IRC | 22:16 | |
| ohnoimdead | ayoung: you around? | 22:16 |
| *** martine has quit IRC | 22:18 | |
| mnewby | any devstack devs around? | 22:18 |
| *** pasik has quit IRC | 22:18 | |
| ohnoimdead | termie: maybe you can answer a keystone+ldap question? | 22:18 |
| *** cp16net has quit IRC | 22:19 | |
| *** johnpostlethwait has quit IRC | 22:19 | |
| termie | ohnoimdead: hahahahahahahhahahahahahahaha | 22:21 |
| termie | ohnoimdead: i mean, sure | 22:21 |
| ohnoimdead | termie: YAY! it's an easy one (probably). What should we pass as email and password for create user in the case of ldap as the auth backend? | 22:22 |
| ohnoimdead | since there's just the one create user call which expects email and password as arguments | 22:22 |
| termie | ohnoimdead: are you actually creating the users from the crud rather than in whatever ldap stuff people use? | 22:23 |
| termie | ohnoimdead: i assumed the user info was being pulled from ldap | 22:23 |
| *** pasik has joined #openstack-dev | 22:23 | |
| dtroyer | mnewby: yo | 22:24 |
| mnewby | dtroyer: | 22:24 |
| mnewby | I've been fighting with a dependency issue in devstack - melance inadvertently installed a newer version of WebOb than most projects were compatible with. | 22:24 |
| mnewby | It got me thinking about the need to maintain version parity across projects. | 22:25 |
| mnewby | Which covers regular deployment. | 22:25 |
| ohnoimdead | termie: so you are saying that create user should never be called when using ldap? i though keystone was keeping some metadata/authz stuff for users? | 22:25 |
| mnewby | But I was thinking that devstack would probably benefit from installing each service in an isolated venv, to ensure that version skew in development doesn't become a problem. | 22:25 |
| mnewby | Thoughts? | 22:25 |
| ohnoimdead | termie: for reference i'm working on: https://bugs.launchpad.net/horizon/+bug/948310 | 22:25 |
| uvirtbot | Launchpad bug 948310 in horizon "User password isn't a required field on creation" [Critical,In progress] | 22:25 |
| *** pixelbeat has quit IRC | 22:26 | |
| dtroyer | mnewby: hmmm…that would probably work for the stuff installed via pip, and theoretically you'd just override the apt packages if required anyway. | 22:27 |
| termie | ohnoimdead: yeah we keep some metadata, but i am not sure how that interaction is supposed to work | 22:27 |
| mnewby | dtroyer: correct | 22:27 |
| Ryan_Lane | ohnoimdead, termie: I'd imagine keystore can add the metadata/authz stuff when pulled from ldap | 22:27 |
| mnewby | dtroyer: it wouldn't be possible to have skew in os packages. | 22:27 |
| Ryan_Lane | or is there some things added upon creation? | 22:27 |
| Ryan_Lane | *are | 22:27 |
| *** kbringard has quit IRC | 22:28 | |
| termie | Ryan_Lane, ohnoimdead: i'm not really sure how it is being done | 22:28 |
| * Ryan_Lane nods | 22:28 | |
| termie | i haven't used it myself, although rpedde has, and ayoung would probably iknwo | 22:28 |
| dtroyer | mnewby: trying to decide if that should be global or per-repo, thoughts? | 22:28 |
| cloudfly | dtroyer i also noticed with the cloudfiles problem back at beginning of month there were untracked python dependencies in /usr/local/lib | 22:28 |
| *** zzed has joined #openstack-dev | 22:28 | |
| ohnoimdead | ok, so it sounds like what we need to do for this bug is simply disable the user create button when keystone is using ldap? | 22:28 |
| cloudfly | which was annoying to isolate | 22:28 |
| mnewby | dtroyer: i imagine each repo getting it's own venv. how would a global install work? | 22:29 |
| dtroyer | mnewby: sorry, I was just thinking about the setting to make avenv or not | 22:29 |
| mnewby | ah, right. | 22:29 |
| mnewby | dtroyer: that makes sense. | 22:29 |
| mnewby | dtroyer: although, which would be the default? | 22:30 |
| dtroyer | mnewby: the default would be like it is now, no vent. We'd add a switch to enable it | 22:30 |
| dtroyer | cloudfly: what did you find in /usr/local/lib? | 22:31 |
| mnewby | dtroyer: I would recommend using a venv by default. otherwise version skew in any given project would effectively break devstack, as I experienced recently. | 22:31 |
| mnewby | dtroyer: but I guess some developer experience would help | 22:31 |
| dtroyer | mnewby: also, at some point the packagers will have to resolve that. do any packages use vents today? | 22:32 |
| mnewby | dtroyer: some guinea pigs (like me) could test out the venv option and report on how it affected development. | 22:32 |
| dtroyer | dang, can't type tailing 'v' today it seems | 22:32 |
| dtroyer | or trailing | 22:32 |
| mnewby | dtroyer: I imagine venvs being only for development, with packaging forcing the versions of dependencies to converge. | 22:32 |
| mnewby | dtroyer; having venv support in devstack would be to allow temporary version skew, not long-term. | 22:33 |
| dtroyer | and that's one reason I would default it to off too. I don't see using vents for gating tests | 22:33 |
| mnewby | dtroyer: so that developers could incrementally upgrade to a new dependency rather than having to do so all at once. | 22:33 |
| mnewby | dtroyer: ah, true. | 22:33 |
| mnewby | dtroyer: So it could be optionally turned on during times of dependency version changes, but otherwise left alone. | 22:34 |
| mnewby | dtroyer: Makes sense. | 22:34 |
| dtroyer | mnewby: sounds like a good idea, can you write up a bug? | 22:34 |
| *** lloydde has quit IRC | 22:34 | |
| mnewby | Can do. | 22:34 |
| dweimer | In swift, what's the reason for the object metadata size limits? | 22:37 |
| *** andrewsmedina has joined #openstack-dev | 22:37 | |
| termie | ohnoimdead: that might be a good solution for now | 22:37 |
| *** markvoelker has quit IRC | 22:38 | |
| *** mattray has quit IRC | 22:41 | |
| *** jakedahn is now known as jakedahn_zz | 22:42 | |
| *** johnpostlethwait has joined #openstack-dev | 22:43 | |
| *** apevec has quit IRC | 22:44 | |
| *** jakedahn_zz is now known as jakedahn | 22:56 | |
| *** dalang has quit IRC | 22:57 | |
| *** johnpostlethwait has quit IRC | 22:57 | |
| *** johnpostlethwait has joined #openstack-dev | 22:57 | |
| *** deshantm has joined #openstack-dev | 23:06 | |
| *** bsza1 has quit IRC | 23:09 | |
| *** mdomsch has quit IRC | 23:15 | |
| vishy | mnewby, dtroyer I like keeping shared install for gating, I think the issue is that melange is not in gating | 23:16 |
| *** pixelbeat has joined #openstack-dev | 23:17 | |
| mnewby | vishy: ah, ok. | 23:18 |
| vishy | mnewby: I would prefer not to have bugs like that exist for even a short time | 23:18 |
| vishy | mnewby: perhaps non-core projects could be installed into a venv by default? | 23:18 |
| mnewby | vishy: the intent isn't to take away shared install though, but rather to allow non-shared install to support updating dependencies incrementally. | 23:18 |
| mnewby | vishy: the default would be a shared install | 23:19 |
| mnewby | vishy: ah, that makes sense too | 23:19 |
| *** dalang has joined #openstack-dev | 23:19 | |
| mnewby | vishy: Anything that's not in gating could be a venv install by default. | 23:19 |
| *** jimbaker has quit IRC | 23:20 | |
| *** dolphm has quit IRC | 23:21 | |
| dtroyer | mnewby, vishy: since this is meant to be for transitional periods, I think it should be an intentional act to handle those exceptions/transitions. I want to reinforce the notion that a venv isn't 'normal' | 23:23 |
| mnewby | dtroyer, vishy: I think it makes sense to include non-gated projects in a venv by default, though, so that those projects can't screw up a devstack install. | 23:24 |
| mnewby | dtroyer, vishy: Otherwise, what's the use in gating if it can't guarantee a working devstack? | 23:25 |
| *** mattray has joined #openstack-dev | 23:27 | |
| dubsquared | mattray: this doesn't look like work | 23:29 |
| *** Glacee has quit IRC | 23:30 | |
| mattray | IRC is an occupational hazard | 23:36 |
| mnewby | jkoelker: Are you here? | 23:40 |
| *** deshantm_ has joined #openstack-dev | 23:40 | |
| *** deshantm has quit IRC | 23:44 | |
| *** jakedahn is now known as jakedahn_zz | 23:47 | |
| *** lloydde has joined #openstack-dev | 23:48 | |
| *** lloydde has quit IRC | 23:48 | |
| *** mattray has quit IRC | 23:48 | |
| notmyname | ttx: swift 1.4.7 is good to go for official release tomorrow | 23:49 |
| *** dubsquared has quit IRC | 23:49 | |
| *** roge has quit IRC | 23:50 | |
| *** AlanClark has quit IRC | 23:50 | |
| *** roge has joined #openstack-dev | 23:51 | |
| *** reed_ has quit IRC | 23:51 | |
| *** reed_ has joined #openstack-dev | 23:52 | |
| *** AlanClark has joined #openstack-dev | 23:52 | |
| *** dubsquared has joined #openstack-dev | 23:52 | |
| *** dtroyer has quit IRC | 23:57 | |
| *** griff has joined #openstack-dev | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!