| *** adjohn has joined #openstack-dev | 00:00 | |
| *** _adjohn has quit IRC | 00:00 | |
| comstud | vishy: i have a client to communicate with the guest agent that does DH without m2crypto or any crypto modules. If it's important to remove it from nova, I can prop something | 00:01 |
|---|---|---|
| *** mikeyp has joined #openstack-dev | 00:02 | |
| *** dtroyer has quit IRC | 00:03 | |
| *** jeffjapan has joined #openstack-dev | 00:21 | |
| openstackgerrit | Brian Waldon proposed a change to openstack/glance: Add policy checking for basic image operations https://review.openstack.org/3130 | 00:23 |
| openstackgerrit | sleepsonthefloor proposed a change to openstack-dev/devstack: Add nova's new vnc support to devstack https://review.openstack.org/3195 | 00:24 |
| *** bencherian has joined #openstack-dev | 00:25 | |
| *** sdake has quit IRC | 00:25 | |
| *** sdake has joined #openstack-dev | 00:26 | |
| *** bcwaldon_ has joined #openstack-dev | 00:31 | |
| *** bcwaldon has quit IRC | 00:31 | |
| *** bcwaldon_ is now known as bcwaldon | 00:31 | |
| *** CarlosM has quit IRC | 00:32 | |
| *** lloydde has quit IRC | 00:37 | |
| *** deshantm has quit IRC | 00:50 | |
| *** shevek_ has joined #openstack-dev | 00:53 | |
| *** bengrue has quit IRC | 01:00 | |
| *** vladimir3p has quit IRC | 01:02 | |
| *** vladimir3p has joined #openstack-dev | 01:08 | |
| *** vladimir3p has quit IRC | 01:09 | |
| *** ayoung has quit IRC | 01:14 | |
| *** jog0 has left #openstack-dev | 01:15 | |
| *** ayoung has joined #openstack-dev | 01:18 | |
| *** jakedahn has quit IRC | 01:18 | |
| *** bencherian has quit IRC | 01:21 | |
| *** sleepsonthefloo has quit IRC | 01:24 | |
| *** dwalleck_nova has joined #openstack-dev | 01:25 | |
| *** dtroyer has joined #openstack-dev | 01:26 | |
| *** pixelbeat has quit IRC | 01:26 | |
| *** lloydde has joined #openstack-dev | 01:30 | |
| *** zns has joined #openstack-dev | 01:30 | |
| *** reed has quit IRC | 01:34 | |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Create nova cert extension for x509 support https://review.openstack.org/3199 | 01:38 |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Adds extension for retrieving certificates https://review.openstack.org/3200 | 01:38 |
| openstackgerrit | A change to openstack/nova has been rejected: Adds extension for retrieving certificates https://review.openstack.org/3200 | 01:39 |
| openstackgerrit | Vish Ishaya proposed a change to openstack-dev/devstack: Add cert worker to devstack https://review.openstack.org/3194 | 01:40 |
| vishy | smoser: got it euca-upload-bundle working again with the above ^^ | 01:41 |
| vishy | smoser: once we have it all in, we can add a button to dashboard allowing people to download their euca credentials | 01:42 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Fix LDAP Schema Syntax (bug 904380) https://review.openstack.org/3140 | 01:42 |
| uvirtbot | Launchpad bug 904380 in keystone ""Duplicate option" error when adding the LDAP schema to OpenLDAP" [Medium,Incomplete] https://launchpad.net/bugs/904380 | 01:42 |
| smoser | vishy, woot! | 01:43 |
| *** andrewbogott has quit IRC | 01:43 | |
| *** crobinso has joined #openstack-dev | 01:43 | |
| openstackgerrit | Verification of a change to openstack/keystone failed: Bug #916199: keystone-manage service list fails with AttributeError on Service.description https://review.openstack.org/3047 | 01:52 |
| uvirtbot | Launchpad bug 916199 in keystone "keystone-manage service list fails with AttributeError on Service.description" [Undecided,In progress] https://launchpad.net/bugs/916199 | 01:52 |
| *** lloydde has quit IRC | 01:52 | |
| *** dtroyer has quit IRC | 01:53 | |
| *** dtroyer has joined #openstack-dev | 01:53 | |
| *** dwalleck_nova has quit IRC | 01:59 | |
| *** lloydde has joined #openstack-dev | 01:59 | |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Pre-staging pip requires https://review.openstack.org/3201 | 01:59 |
| *** maplebed has quit IRC | 02:00 | |
| *** CarlosM has joined #openstack-dev | 02:02 | |
| zns | ttx, mtaylor, jeblair: anyone here? Anyone that can help with the OpenStack Jenkins servers? | 02:02 |
| *** CarlosM1 has joined #openstack-dev | 02:03 | |
| *** CarlosM has quit IRC | 02:06 | |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Fixes 918535: time not properly parsed in auth_token middleware https://review.openstack.org/3185 | 02:11 |
| *** dwalleck_nova has joined #openstack-dev | 02:16 | |
| *** jdurgin has quit IRC | 02:17 | |
| *** misheska has quit IRC | 02:21 | |
| *** bengrue has joined #openstack-dev | 02:24 | |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Prestage fix - fixed requirement name; python-dateutil, not dateutil https://review.openstack.org/3203 | 02:24 |
| *** dwalleck_nova has quit IRC | 02:25 | |
| *** davlap has joined #openstack-dev | 02:27 | |
| *** davlap has quit IRC | 02:28 | |
| openstackgerrit | Verification of a change to openstack/keystone failed: Fixes 918535: time not properly parsed in auth_token middleware https://review.openstack.org/3185 | 02:30 |
| *** CarlosM1 has quit IRC | 02:33 | |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Fixes 918535: time not properly parsed in auth_token middleware https://review.openstack.org/3185 | 02:38 |
| *** Lumiere has quit IRC | 02:38 | |
| *** Lumiere has joined #openstack-dev | 02:39 | |
| openstackgerrit | Verification of a change to openstack/keystone failed: Prestage fix - fixed requirement name; python-dateutil, not dateutil https://review.openstack.org/3203 | 02:39 |
| *** zns has quit IRC | 02:41 | |
| *** zns has joined #openstack-dev | 02:41 | |
| *** Ryan_Lane has quit IRC | 02:47 | |
| *** gyee has quit IRC | 02:47 | |
| *** hugokuo has joined #openstack-dev | 02:48 | |
| *** rajaram has joined #openstack-dev | 02:53 | |
| *** crobinso has quit IRC | 02:56 | |
| *** adjohn has quit IRC | 03:00 | |
| *** shevek_ has quit IRC | 03:07 | |
| mtaylor | zns: sorry - just got back online. looking now | 03:16 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Fixes 918535: time not properly parsed in auth_token middleware https://review.openstack.org/3185 | 03:20 |
| zns | mtaylor: np. thanks for looking. | 03:21 |
| zns | just would like to get this fixed (or have a workaround) before we go into the weekend to get these in for next week. | 03:22 |
| openstackgerrit | Chuck Short proposed a change to openstack/nova: Fixes LP: #919034 https://review.openstack.org/3204 | 03:31 |
| mtaylor | zns: ++ | 03:33 |
| mtaylor | zns: still trying to figure out why it's borking - I've put in some sleeps to see if it's a concurrency issue | 03:33 |
| mtaylor | zns: looks like it's a thread pool issue ... the sleeps I put in shoud alleviate it for now, we'll get the config fixed asap | 03:38 |
| zns | mtaylor: tx. I'll push a few in now and see how it goes... | 03:38 |
| mtaylor | zns: thanks! | 03:38 |
| mtaylor | zns: and sorry it's been kicking you today | 03:39 |
| *** Ryan_Lane has joined #openstack-dev | 03:42 | |
| zaitcev | I have a silly question, guys. How is this made: http://launchpad.net/glance/diablo/2011.3.1/+download/glance-2011.3.1.tar.gz | 03:44 |
| *** andrewsmedina has joined #openstack-dev | 03:45 | |
| zaitcev | For example, in typical projects with GNU toolchain and autoconf, one runs "make dist" and this produces the tarball. And in our case? | 03:45 |
| zaitcev | Does setuptools do it, and it yes, how? | 03:47 |
| *** andrewsmedina has quit IRC | 03:47 | |
| *** dolphm has joined #openstack-dev | 03:48 | |
| openstackgerrit | Verification of a change to openstack/keystone failed: Fixes 918535: time not properly parsed in auth_token middleware https://review.openstack.org/3185 | 03:48 |
| *** Ryan_Lane1 has joined #openstack-dev | 03:50 | |
| *** mdomsch has joined #openstack-dev | 03:51 | |
| zns | mtaylor: still happening :-( https://jenkins.openstack.org/job/gate-keystone-unittests/280/console | 03:52 |
| mtaylor | zns: that makes me sad - but thanks | 03:52 |
| *** Ryan_Lane has quit IRC | 03:54 | |
| openstackgerrit | Verification of a change to openstack/keystone failed: fix bug lp:843064 https://review.openstack.org/2835 | 03:58 |
| *** jog0 has joined #openstack-dev | 04:03 | |
| *** adjohn has joined #openstack-dev | 04:03 | |
| mtaylor | zns: ok. the workaround I put in was, well, not a good one. :) according to the thread I emailed you, it's potentially an https vs. ssh behavior issue, combined with a queue length issue | 04:05 |
| *** dwalleck_nova has joined #openstack-dev | 04:05 | |
| mtaylor | zns: for now, we've set jenkins to use ssh rather than https to pull the changes, so that at least we won't get disconnected quickly (which the https handler is doing currently0- | 04:06 |
| *** jog0 has quit IRC | 04:06 | |
| mtaylor | zns: we're also looking at gerrit tuning parameters to help deal with the internal operation queue | 04:06 |
| zns | mtaylor: OK. Will try again. | 04:15 |
| mtaylor | zns: I resubmitted that one change | 04:15 |
| mtaylor | zns: we've isolated the cause - mainly working on solid solutions long term atm | 04:15 |
| openstackgerrit | Verification of a change to openstack/keystone failed: fix bug lp:843064 https://review.openstack.org/2835 | 04:16 |
| zaitcev | Aaaw screw it all. I'll just run diff -ur -X dontdiff ../rpms/BUILD/glance-2011.3 glance-wk, then apply as a patch with rpmbuild, FTW. | 04:16 |
| mtaylor | zaitcev: sorry - didn't see your question before | 04:17 |
| mtaylor | zaitcev: the command is "python setup.py sdist" | 04:17 |
| mtaylor | zaitcev: and then the output goes in to the dist dir | 04:18 |
| mtaylor | zaitcev: as a tarball | 04:18 |
| zaitcev | mtaylor: thanks! | 04:18 |
| zns | mtaylor: can you bump the timeout up to 15 minutes for https://jenkins.openstack.org/job/gate-keystone-unittests/configure? | 04:20 |
| mtaylor | zns: I can - are we hitting it? | 04:22 |
| mtaylor | zns: AH - I se it | 04:22 |
| mtaylor | zns: done | 04:23 |
| mtaylor | zns: and re-triggered | 04:23 |
| zns | mtaylor: we'll put some focus on speeding things up once we're done getting features in for E3... | 04:23 |
| zns | tx | 04:23 |
| zns | mtaylor: dunno if I missed your response earlier. Is there a way to retrigger one failed job and pass the review or do we have to go back and re-approve in Gerrit? | 04:24 |
| *** dolphm has quit IRC | 04:25 | |
| mtaylor | zns: right now re-approve is the easiest way - I need to get a good way to give approvers re-trigger rights directly in jenkins | 04:25 |
| zns | ok | 04:25 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Fixes 918535: time not properly parsed in auth_token middleware https://review.openstack.org/3185 | 04:27 |
| *** dolphm has joined #openstack-dev | 04:31 | |
| *** sleepsonthefloo has joined #openstack-dev | 04:32 | |
| *** andrewsmedina has joined #openstack-dev | 04:34 | |
| *** dolphm has quit IRC | 04:50 | |
| openstackgerrit | James E. Blair proposed a change to openstack/openstack-ci-puppet: Tune gerrit parameters. https://review.openstack.org/3206 | 04:50 |
| jeblair | mtaylor: ^ | 04:50 |
| mtaylor | jeblair: thanks! | 04:51 |
| *** shevek_ has joined #openstack-dev | 04:52 | |
| *** andrewsmedina has quit IRC | 04:52 | |
| vishy | comstud: ping | 04:57 |
| comstud | vishy: pong | 04:58 |
| vishy | comstud: hey, we are hacking on scheduler stuff, making some new filters | 04:58 |
| comstud | vishy: Ok. | 04:58 |
| vishy | comstud: unfortunately we have no context in the filter | 04:58 |
| vishy | comstud: so I can't do special filtering for admins | 04:58 |
| comstud | vishy: I do have some other minor changes coming in an hour or 2 | 04:58 |
| vishy | comstud: do you think I should pass the context? | 04:58 |
| comstud | vishy: Ah, perhaps store it in HostState? | 04:59 |
| comstud | in my private branch, I add other things like 'project_id' | 04:59 |
| comstud | but wasn't sure that was useful for trunk.. maybe it is | 04:59 |
| vishy | comstud: it really is part of the schedule request | 04:59 |
| vishy | I'm worried about shared state | 04:59 |
| comstud | passing as an explicit argument makes sense to me then | 05:00 |
| vishy | is there some object that is unique to the schedule request? | 05:00 |
| vishy | that I can store it in? | 05:00 |
| comstud | filter_properties | 05:00 |
| comstud | if you look at distributed scheduler | 05:00 |
| comstud | look at what we build right now for filter_properties | 05:00 |
| comstud | you probably want to do something similar | 05:00 |
| comstud | you could add context or project_id in there | 05:00 |
| comstud | or whatever you need | 05:00 |
| comstud | that gets passed all of the way down to the filters | 05:01 |
| vishy | comstud: good idea, I will pass it in properties | 05:02 |
| comstud | that's essentially its purpose | 05:02 |
| comstud | btw, i have a patch coming that'll touch things slightly | 05:02 |
| comstud | 'service' is not set with HostState and it's needed | 05:02 |
| comstud | i'm moving service_is_up() from driver.Scheduler into utils | 05:02 |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Fixes 918535: time not properly parsed in auth_token middleware https://review.openstack.org/3185 | 05:03 |
| comstud | because ec2 API essentially does the same check | 05:03 |
| comstud | and we need to be able to use service_is_up in the filters. | 05:03 |
| comstud | i'm working on the tests right now, then it'll be propped | 05:03 |
| _0x44 | https://gist.github.com/c25ccfa2eea0418b46f7 | 05:11 |
| *** jog0 has joined #openstack-dev | 05:11 | |
| *** dwalleck_nova has quit IRC | 05:14 | |
| openstackgerrit | deepak garg proposed a change to openstack/nova: bp/xenapi-provider-firewall and bug/915403 https://review.openstack.org/3207 | 05:15 |
| vishy | comstud: cool | 05:16 |
| jog0 | comstud, Vishy tells me you have been working on the scheduler | 05:16 |
| comstud | jog0: yep | 05:16 |
| vishy | comstud, all: here is where I am so far https://github.com/vishvananda/nova/compare/master...scheduler-clean | 05:17 |
| vishy | adding an availability zone filter next | 05:17 |
| jog0 | whats the logic behind allowing the filters (default_host_filters flag) being overridden by a scheduler? | 05:17 |
| openstackjenkins | Project nova-docs build #1986: SUCCESS in 2 min 54 sec: https://jenkins.openstack.org/job/nova-docs/1986/ | 05:18 |
| openstackjenkins | jogo: aws/ec2 api validation | 05:18 |
| comstud | missing 'context' added to 'def filter_properties()' in dist scheduler | 05:19 |
| comstud | but that makes sense | 05:19 |
| jog0 | instead of just having the chosen filters set by a flag (and resulting in it being set in nova.conf | 05:19 |
| jog0 | instead of hiding what filters are being used inside of the scheduler | 05:19 |
| *** novas0x2a|laptop has quit IRC | 05:20 | |
| comstud | jog0: We're going to want the ability to be able to modify filters and rules and weights and so forth without having to restart the service... there's a json options file that's currently unused. i suspect we'll want to use that to add support to be able to specify filter rules, etc, at some point. You probably notice we're not using the ability to override the default rules right now | 05:20 |
| *** novas0x2a|laptop has joined #openstack-dev | 05:20 | |
| comstud | jog0: That sorta lives because at one point we might want filters to come in with a build request... | 05:21 |
| vishy | comstud: where in filter_properties might i find instance parameters? | 05:21 |
| comstud | perhaps in a bursting architecture | 05:21 |
| vishy | like which availability zone was specified? | 05:21 |
| comstud | vishy: They'll need added | 05:21 |
| comstud | rather than shove a bunch of stuff in there, I left it mostly empty for now in dist scheduler | 05:22 |
| comstud | in the dist scheduler... | 05:22 |
| vishy | comstud: it doesn't pass in everything from the compute create command? | 05:22 |
| vishy | like instance_opts dict? | 05:22 |
| comstud | oh | 05:23 |
| comstud | nevermind, it's there | 05:23 |
| jog0 | comstud, thanks so its fur the future | 05:23 |
| comstud | request_spec['instance_properties'] | 05:23 |
| vishy | ok cool | 05:23 |
| vishy | comstud: and stuff from the services table is in host_state.service? | 05:23 |
| comstud | i suppose that gives you the project_id too | 05:23 |
| comstud | vishy: It will be, yes | 05:23 |
| comstud | almost done with that | 05:24 |
| vishy | seems like we might as well pass the entire request_spec into the filter | 05:25 |
| vishy | but for now I will add availability_zone from it manually | 05:25 |
| comstud | i thought we did | 05:25 |
| comstud | maybe not | 05:25 |
| comstud | i guess not | 05:25 |
| *** jog0 has quit IRC | 05:25 | |
| *** jog0 has joined #openstack-dev | 05:26 | |
| comstud | i agree with adding it | 05:26 |
| *** jog0_ has joined #openstack-dev | 05:27 | |
| *** jog0 has quit IRC | 05:27 | |
| *** jog0_ is now known as jog0 | 05:27 | |
| vishy | should i just do a merge? or stick it in filter_properties['request_spec'] ? | 05:27 |
| comstud | i would put it as filter_properties['request_spec'] | 05:27 |
| vishy | k | 05:28 |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Implemented bp keystone-manage2 https://review.openstack.org/3123 | 05:29 |
| *** novas0x2a|laptop has quit IRC | 05:31 | |
| comstud | there's a bug in the json filter with 'not'.. just going to ignore it | 05:33 |
| openstackgerrit | Chris Behrens proposed a change to openstack/nova: scheduler host_manager needs service for filters https://review.openstack.org/3208 | 05:38 |
| *** zns1 has joined #openstack-dev | 05:40 | |
| *** jog0 has left #openstack-dev | 05:41 | |
| *** jog0 has joined #openstack-dev | 05:41 | |
| *** zns has quit IRC | 05:42 | |
| openstackgerrit | Chris Behrens proposed a change to openstack/nova: scheduler host_manager needs service for filters https://review.openstack.org/3208 | 05:44 |
| vishy | comstud: nice looks good | 05:45 |
| vishy | comstud: my az filter is here: https://github.com/vishvananda/nova/commit/6ee798c1301bc231cac09681188cf966c8c26968 | 05:45 |
| comstud | still cleaning up stuff i see in the diff | 05:46 |
| openstackgerrit | Chris Behrens proposed a change to openstack/nova: scheduler host_manager needs service for filters https://review.openstack.org/3208 | 05:46 |
| comstud | alright | 05:47 |
| comstud | that should do it | 05:47 |
| comstud | dangit, left an unused import | 05:47 |
| *** jaypipes-afk has quit IRC | 05:55 | |
| *** jaypipes-afk has joined #openstack-dev | 05:55 | |
| comstud | i need to get my scheduler refactor branch up too | 05:57 |
| comstud | scheduler tests take too long | 05:57 |
| comstud | er | 05:57 |
| comstud | scheduler unit tests | 05:57 |
| comstud | branch | 05:57 |
| comstud | cuts from 30-40s -> 3s | 05:57 |
| comstud | or something. | 05:58 |
| openstackgerrit | justinsb proposed a change to openstack/nova: Support filter based on CPU core (over)allocation https://review.openstack.org/3209 | 05:58 |
| _0x44 | vishy: https://gist.github.com/c25ccfa2eea0418b46f7 | 05:58 |
| comstud | ah, Bro filters. | 05:59 |
| openstackgerrit | Chris Behrens proposed a change to openstack/nova: scheduler host_manager needs service for filters https://review.openstack.org/3208 | 06:00 |
| vishy | hehe | 06:00 |
| comstud | _0x44, vishy: if you guys can take 5 minutes to review ^^ | 06:00 |
| comstud | then you can merge | 06:00 |
| bcwaldon | no | 06:00 |
| comstud | or waldon | 06:00 |
| bcwaldon | not a chance | 06:00 |
| comstud | he doesn't do anything | 06:00 |
| comstud | come on man | 06:00 |
| comstud | <3 | 06:00 |
| openstackgerrit | Chris Behrens proposed a change to openstack/nova: scheduler host_manager needs service for filters https://review.openstack.org/3208 | 06:02 |
| comstud | the longer you guys wait, the more i change it. | 06:02 |
| _0x44 | Dammit, stop doing that. | 06:03 |
| comstud | i'm done | 06:03 |
| _0x44 | I had just hit the review button | 06:03 |
| comstud | yeah | 06:03 |
| comstud | it's funny | 06:03 |
| comstud | i got an email about a comment you made | 06:03 |
| comstud | that was blank | 06:04 |
| comstud | and i look at the review and it's not there | 06:04 |
| vishy | dammit comstud | 06:04 |
| comstud | i re-added a test i accidentally removed.. sue me | 06:05 |
| comstud | :) | 06:05 |
| comstud | thnx! | 06:05 |
| openstackgerrit | Zhongyue Luo proposed a change to openstack/nova: nova flags list tool https://review.openstack.org/3210 | 06:12 |
| *** Ryan_Lane1 has quit IRC | 06:12 | |
| comstud | dropping off for a bit | 06:13 |
| comstud | bbiab, you kids have fun. | 06:13 |
| openstackgerrit | Joe Gordon proposed a change to openstack/nova: Pull out ram_filter into a separate filter https://review.openstack.org/3211 | 06:18 |
| *** zaitcev has quit IRC | 06:19 | |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Implemented bp keystone-manage2 https://review.openstack.org/3123 | 06:20 |
| openstackgerrit | Christopher MacGown proposed a change to openstack/nova: Add simple affinity filters, bro https://review.openstack.org/3212 | 06:20 |
| openstackgerrit | Christopher MacGown proposed a change to openstack/nova: Add simple affinity filters, bro https://review.openstack.org/3212 | 06:23 |
| comstud | i'm really not sure I can approve filters with 'Bro' in them. | 06:23 |
| openstackgerrit | Verification of a change to openstack/nova failed: scheduler host_manager needs service for filters https://review.openstack.org/3208 | 06:25 |
| comstud | whhat | 06:25 |
| comstud | dev stack, come on | 06:25 |
| comstud | Build timed out (after 20 minutes). Marking the build as failed. | 06:27 |
| comstud | hmm. | 06:27 |
| *** sleepsonthefloo has quit IRC | 06:30 | |
| openstackgerrit | Christopher MacGown proposed a change to openstack/nova: Add simple affinity filters, bro https://review.openstack.org/3212 | 06:31 |
| comstud | vishy: Is there any reason checking service['disabled'] in simple scheduler would cause devstack to fail? | 06:31 |
| vishy | comstud: hmm, perhaps | 06:32 |
| comstud | vishy: Why would you have a service set to disabled? | 06:32 |
| comstud | or is it somehow NULL in the db? | 06:32 |
| vishy | comstud: depends on if they start disabled by default | 06:32 |
| vishy | comstud: there is a flag iirc | 06:32 |
| comstud | hm i don't think they start disabled | 06:32 |
| vishy | comstud: oh | 06:32 |
| vishy | comstud: simple scheduler already checked the disabled flag | 06:33 |
| vishy | comstud: i think | 06:33 |
| comstud | nope | 06:33 |
| comstud | it checked service_is_up() | 06:33 |
| vishy | comstud: possible it is running before report_state has updated stuff? | 06:33 |
| comstud | which only checks the down_time | 06:33 |
| comstud | this comes from the DB | 06:33 |
| comstud | services table | 06:33 |
| comstud | db.service_get* | 06:34 |
| vishy | comstud: disabled == 0 for all of mine | 06:34 |
| vishy | in devstack | 06:34 |
| comstud | then that should be okay | 06:34 |
| comstud | 83 if utils.service_is_up(service) and not service['disabled']: 84 return service['host'] | 06:35 |
| comstud | i added that 'disabled' check there | 06:35 |
| comstud | so that should still be true | 06:35 |
| comstud | so, i see no other reason why devstack would fail with a build timeout | 06:35 |
| comstud | unless it's an environmental issue | 06:35 |
| comstud | apparently I need to start using devstack | 06:36 |
| comstud | i'm going to try pushing this again to see if it was transient | 06:37 |
| *** rmkdroid has joined #openstack-dev | 06:39 | |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Fixed PEP8 violations and disallowed them https://review.openstack.org/3189 | 06:39 |
| openstackgerrit | justinsb proposed a change to openstack/nova: Support filter based on CPU core (over)allocation https://review.openstack.org/3209 | 06:47 |
| openstackgerrit | Verification of a change to openstack/nova failed: scheduler host_manager needs service for filters https://review.openstack.org/3208 | 06:58 |
| vishy | error: RPC failed; result=22, HTTP code = 503 | 06:58 |
| vishy | comstud: ^^ | 06:58 |
| comstud | what is RPC in that case? | 06:58 |
| openstackgerrit | Brian Waldon proposed a change to openstack/python-novaclient: Match create_image on server object and manager https://review.openstack.org/3213 | 07:00 |
| vishy | it is gerrit failing to download your changeset... | 07:01 |
| vishy | :o | 07:01 |
| comstud | interesting | 07:01 |
| comstud | i saw the git cmd right before it | 07:01 |
| comstud | but wasn't sure | 07:01 |
| *** Ryan_Lane has joined #openstack-dev | 07:02 | |
| vishy | at least, that is how i read it | 07:03 |
| comstud | well, that puts a cramp in the hackathon! | 07:03 |
| vishy | comstud: got my az filter working | 07:04 |
| comstud | great! | 07:04 |
| vishy | we are working on pulling all of the checks into compute into separate filters | 07:04 |
| vishy | the ram filter just bit me in the ass | 07:04 |
| vishy | would be nice to be able to turn it off | 07:04 |
| comstud | you could separate it into its own filter class | 07:05 |
| vishy | comstud: yup that is what we are doing | 07:05 |
| comstud | perhaps we should bea ble to list individual methods to run | 07:05 |
| comstud | instead of classes | 07:05 |
| comstud | so you don't have to have a million classes | 07:05 |
| vishy | comstud: interesting thought. I don't mind the class based version too much right now | 07:06 |
| comstud | go with it for now | 07:06 |
| vishy | mtaylor, jeblair: either of you here? | 07:06 |
| comstud | i'm curious what case you don't want to check ram | 07:06 |
| comstud | ? | 07:06 |
| comstud | i can think of overcommiting as being desired on an appropriate hypervisor that supports it | 07:07 |
| vishy | my devstack vm doesn't have enough free ram to allocate 512 | 07:07 |
| vishy | which means i can't test a vm launch | 07:07 |
| vishy | so basically just turning it off for dev... | 07:07 |
| comstud | ahhh | 07:08 |
| comstud | you could just set the compute_node table appropriately | 07:08 |
| vishy | comstud: fetch worked this time so I would expect your branch to merge | 07:08 |
| comstud | ok | 07:08 |
| comstud | gotta drop off | 07:09 |
| comstud | hot tub time | 07:09 |
| * comstud & bbiab | 07:09 | |
| *** hugokuo has quit IRC | 07:14 | |
| *** adjohn has quit IRC | 07:16 | |
| *** lloydde has quit IRC | 07:23 | |
| jog0 | heres an early version of our boto/EC2 stress tester for nova https://github.com/cloudscaling/tarkin | 07:24 |
| *** anotherjesse has quit IRC | 07:24 | |
| * ttx yawns | 07:25 | |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Adds Availability Zone Filter https://review.openstack.org/3214 | 07:30 |
| *** cp16net_ has joined #openstack-dev | 07:33 | |
| openstackjenkins | Project nova-docs build #1987: SUCCESS in 2 min 37 sec: https://jenkins.openstack.org/job/nova-docs/1987/ | 07:33 |
| openstackjenkins | Chris Behrens: scheduler host_manager needs service for filters | 07:33 |
| *** LinuxJedi has joined #openstack-dev | 07:34 | |
| *** cp16net has quit IRC | 07:36 | |
| openstackgerrit | Christopher MacGown proposed a change to openstack/nova: Add simple affinity filters, bro Update the filter so that context is part of the filter_properties, bro. Add todo to change instance_id get when filter_properties is solidified https://review.openstack.org/3215 | 07:36 |
| *** jog0 has quit IRC | 07:36 | |
| *** Ryan_Lane has quit IRC | 07:48 | |
| *** bepernoot has joined #openstack-dev | 07:56 | |
| *** bepernoot has joined #openstack-dev | 07:56 | |
| *** n0ano has quit IRC | 07:58 | |
| *** jog0 has joined #openstack-dev | 07:59 | |
| *** sdake has quit IRC | 08:00 | |
| *** jog0 has quit IRC | 08:00 | |
| *** jog0 has joined #openstack-dev | 08:10 | |
| *** rmkdroid has quit IRC | 08:14 | |
| *** reidrac has joined #openstack-dev | 08:17 | |
| *** jog0 has left #openstack-dev | 08:20 | |
| *** rmkdroid has joined #openstack-dev | 08:20 | |
| *** corrigac has joined #openstack-dev | 08:31 | |
| openstackgerrit | Gabriel Hurley proposed a change to openstack/horizon: Auto-remove auto-generated .rst files for deleted .py files. https://review.openstack.org/3216 | 08:33 |
| openstackgerrit | Gabriel Hurley proposed a change to openstack/horizon: Added a contributing guide to the Horizon docs. https://review.openstack.org/3217 | 08:33 |
| *** ccorrigan has quit IRC | 08:34 | |
| openstackgerrit | Gabriel Hurley proposed a change to openstack/horizon: Tooltip implementation cleanup. https://review.openstack.org/3218 | 08:37 |
| openstackgerrit | Hengqing Hu proposed a change to openstack/nova: Respect availability_zone in nova-volume https://review.openstack.org/2956 | 08:50 |
| *** pixelbeat has joined #openstack-dev | 08:51 | |
| openstackgerrit | Joe Gordon proposed a change to openstack/nova: Add HACKING compliance testing to run_test.sh https://review.openstack.org/3219 | 08:52 |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Adds Availability Zone Filter https://review.openstack.org/3214 | 08:57 |
| vishy | comstud: did we ever fix schedule_start_instance ? | 08:59 |
| vishy | comstud: I'm still not sure why it would need to go through the scheduler at all | 08:59 |
| comstud | vishy: Because the instance has already been built | 09:02 |
| comstud | it's just shut down | 09:02 |
| comstud | and lives on a spefific host | 09:03 |
| comstud | Although i guess that's not quite true. | 09:03 |
| comstud | hmm | 09:03 |
| *** mikal has quit IRC | 09:03 | |
| comstud | the host must still be assigned, even though it's destroyed from virt layer | 09:03 |
| comstud | which means we do account for its resources when scheduling | 09:04 |
| comstud | and starting it back up on the same host is ok | 09:04 |
| comstud | i think it could make sense to completely re-schedule it | 09:05 |
| comstud | though it doesn't appear necessary right now | 09:05 |
| *** mikal has joined #openstack-dev | 09:07 | |
| comstud | looks like we never unassign instance['host'] | 09:11 |
| comstud | afaict | 09:11 |
| comstud | (even for deletes, which I guess can be useful for debugging) | 09:12 |
| *** mikal has quit IRC | 09:13 | |
| *** jeffjapan has quit IRC | 09:13 | |
| *** Mandell has quit IRC | 09:14 | |
| *** derekh has joined #openstack-dev | 09:15 | |
| *** mikal has joined #openstack-dev | 09:17 | |
| comstud | vishy: ^ | 09:17 |
| *** mikal has quit IRC | 09:22 | |
| *** mikal has joined #openstack-dev | 09:26 | |
| *** mikal has quit IRC | 09:32 | |
| *** mikal has joined #openstack-dev | 09:36 | |
| *** rmkdroid has quit IRC | 09:39 | |
| *** maploin has joined #openstack-dev | 09:40 | |
| *** maploin has quit IRC | 09:40 | |
| *** maploin has joined #openstack-dev | 09:40 | |
| *** darraghb has joined #openstack-dev | 09:52 | |
| *** rajaram has left #openstack-dev | 09:54 | |
| *** rajaram has joined #openstack-dev | 09:55 | |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Handle EC2 Credentials on /tokens https://review.openstack.org/3220 | 10:13 |
| openstackgerrit | Thierry Carrez proposed a change to openstack/nova: Add a specific filter for kill commands https://review.openstack.org/3142 | 10:53 |
| openstackgerrit | kavan-patil proposed a change to openstack/tempest: Added Keypair extension (os-keypairs) client and tests LP#900139 https://review.openstack.org/3179 | 11:09 |
| *** bepernoot has quit IRC | 11:12 | |
| *** Tribaal has joined #openstack-dev | 11:20 | |
| *** tryggvil_ has joined #openstack-dev | 11:35 | |
| *** bsza has joined #openstack-dev | 11:44 | |
| *** markvoelker has joined #openstack-dev | 12:01 | |
| openstackgerrit | Jenkins proposed a change to openstack/nova: Imported Translations from Launchpad https://review.openstack.org/3221 | 12:33 |
| openstackgerrit | Rajaram Mallya proposed a change to openstack/melange: fixes bug 919055 allowing interfaces to be retrieved without device id https://review.openstack.org/3222 | 12:48 |
| uvirtbot | Launchpad bug 919055 in melange "exception on interface create" [Medium,New] https://launchpad.net/bugs/919055 | 12:48 |
| *** zigo has joined #openstack-dev | 12:50 | |
| openstackgerrit | Rajaram Mallya proposed a change to openstack/melange: fixes bug 919155 by tracking deallocated macs in the allocatable_macs table https://review.openstack.org/3223 | 13:19 |
| uvirtbot | Launchpad bug 919155 in melange "mac addressess are not being reallocated" [Medium,New] https://launchpad.net/bugs/919155 | 13:19 |
| *** rajaram has quit IRC | 13:19 | |
| *** misheska has joined #openstack-dev | 13:22 | |
| openstackgerrit | Sandy Walsh proposed a change to openstack/nova: ComputeNode Capacity support https://review.openstack.org/2460 | 13:47 |
| *** rods has joined #openstack-dev | 13:53 | |
| *** rods has quit IRC | 14:01 | |
| *** dprince has joined #openstack-dev | 14:01 | |
| *** rods has joined #openstack-dev | 14:02 | |
| *** sdake has joined #openstack-dev | 14:02 | |
| smoser | anyone know where this thread ended up: http://www.mail-archive.com/openstack@lists.launchpad.net/msg06148.html | 14:16 |
| smoser | i didn't see a clear conclusion unfortunately. | 14:16 |
| *** lts has joined #openstack-dev | 14:21 | |
| *** misheska has quit IRC | 14:34 | |
| *** rajaram has joined #openstack-dev | 14:37 | |
| *** betodalas_ has joined #openstack-dev | 14:38 | |
| betodalas_ | Hello all, does anybody have installed Windows in Xen with Openstack? | 14:39 |
| *** kbringard has joined #openstack-dev | 14:40 | |
| *** jaypipes-afk is now known as jaypipes | 14:42 | |
| betodalas_ | Hello | 14:54 |
| *** sgd has joined #openstack-dev | 14:56 | |
| ttx | smoser: let me see | 14:58 |
| ttx | smoser: ISTR you all violently agreed, and just waited for someone to actually do the work ? Maybe vishy has another conclusion. | 15:00 |
| ttx | it was a bit derailed by the DHCP vs. stupid company policies discussion. | 15:01 |
| smoser | yeah, so is anyone working on that ? | 15:02 |
| smoser | i'd like to make cloud-init do whatever it is its supposed to do | 15:02 |
| smoser | but i'm not sure what that is | 15:02 |
| ttx | smoser: not that I know of. You should confirm with Vish | 15:03 |
| smoser | vishy, | 15:03 |
| *** benner has quit IRC | 15:05 | |
| *** benner has joined #openstack-dev | 15:06 | |
| *** dtroyer has quit IRC | 15:06 | |
| smoser | jaypipes, around ? | 15:13 |
| jaypipes | smoser: briefly before I run and get coffee :) whatup? | 15:13 |
| smoser | jaypipes, wanted to talk about the "glance.ubuntu.com" thingy | 15:14 |
| smoser | you can get coffee and ping me back | 15:14 |
| jaypipes | smoser: :) np! | 15:14 |
| jaypipes | smoser: how about giving me a call around 11:15? I'll /query you my cell | 15:15 |
| smoser | k | 15:16 |
| *** ahale has quit IRC | 15:17 | |
| *** CarlosM has joined #openstack-dev | 15:20 | |
| *** dtroyer has joined #openstack-dev | 15:20 | |
| *** ahale has joined #openstack-dev | 15:23 | |
| *** zzed has joined #openstack-dev | 15:23 | |
| *** rods has quit IRC | 15:24 | |
| openstackgerrit | Dolph Mathews proposed a change to openstack/keystone: Implemented subparsers (bp keystone-manage2) https://review.openstack.org/3224 | 15:26 |
| *** zns has joined #openstack-dev | 15:27 | |
| *** zns1 has quit IRC | 15:30 | |
| *** zns has quit IRC | 15:30 | |
| *** sgd has quit IRC | 15:32 | |
| *** zns has joined #openstack-dev | 15:33 | |
| *** andrewsmedina has joined #openstack-dev | 15:41 | |
| *** dubsquared has joined #openstack-dev | 15:46 | |
| openstackgerrit | Dan Prince proposed a change to openstack/glance: Set url's on AuthBadRequest exceptions. https://review.openstack.org/3225 | 16:17 |
| openstackgerrit | Tres Henry proposed a change to openstack/horizon: Adds ID to flavors list. https://review.openstack.org/3226 | 16:18 |
| vishy | comstud: that we don't need to reschedule and can just go directly to the host | 16:19 |
| vishy | comstud: but we don't | 16:19 |
| vishy | comstud: 935 FLAGS.scheduler_topic, | 16:19 |
| vishy | 936 {"method": "start_instance", | 16:19 |
| vishy | comstud: so sounds like that is a bug | 16:19 |
| vishy | smoser: let me ask jesse about the disk config stuff | 16:20 |
| vishy | smoser: I will get back to you later | 16:20 |
| betodalas_ | Hello all, does anybody have installed Windows in Xen with Openstack? | 16:21 |
| *** cp16net has joined #openstack-dev | 16:22 | |
| *** lloydde has joined #openstack-dev | 16:24 | |
| *** lloydde has quit IRC | 16:28 | |
| *** adjohn has joined #openstack-dev | 16:29 | |
| comstud | vishy: oh | 16:29 |
| comstud | vishy: Looks like I misread the other day | 16:29 |
| comstud | vishy: I think I saw the FLAGS.compute_topic on 937 when I looked quickly :) | 16:30 |
| _0x44 | comstud: What would you rather the filters be called? | 16:34 |
| comstud | :) | 16:34 |
| comstud | something more professional would be nice | 16:34 |
| comstud | but I'd be willing to just abstain if you don't really want to change it :) | 16:35 |
| _0x44 | comstud: They're not exposed to the requestor are they? It's just something that the ops guys would see, right? | 16:35 |
| _0x44 | comstud: I'm not really invested in the names. | 16:36 |
| comstud | it's something anyone can see, because this is open source | 16:36 |
| comstud | :) | 16:36 |
| *** rnirmal has joined #openstack-dev | 16:37 | |
| _0x44 | Bah, that's not important :) | 16:37 |
| comstud | don't worry, i still love you | 16:38 |
| _0x44 | comstud: BehrensStillLovesMeBroFilter? ;) | 16:45 |
| comstud | if you remove 'Bro', it's good. | 16:45 |
| _0x44 | affinity_filter.ComeAtMeFilter is fine? | 16:46 |
| *** Mandell has joined #openstack-dev | 16:50 | |
| *** deshantm has joined #openstack-dev | 16:51 | |
| *** rnirmal_ has joined #openstack-dev | 16:53 | |
| openstackgerrit | Christopher MacGown proposed a change to openstack/nova: Add simple affinity filters, bro Update the filter so that context is part of the filter_properties, bro. Add todo to change instance_id get when filter_properties is solidified https://review.openstack.org/3215 | 16:53 |
| johan_-_ | who would i talk to about problems with the devstack integration tests in jenkins? | 16:54 |
| johan_-_ | i guess jeblair? | 16:55 |
| *** markmc has quit IRC | 16:55 | |
| *** rnirmal has quit IRC | 16:55 | |
| *** rnirmal_ is now known as rnirmal | 16:55 | |
| openstackgerrit | Christopher MacGown proposed a change to openstack/nova: Add simple affinity filters, bro Update the filter so that context is part of the filter_properties, bro. Remove the Bro, get instance_id out of filter_properties['request_spec']['instance_properties'] https://review.openstack.org/3215 | 16:55 |
| *** adjohn has quit IRC | 16:57 | |
| *** reidrac has quit IRC | 16:58 | |
| *** maploin has quit IRC | 16:59 | |
| *** rmkdroid has joined #openstack-dev | 17:00 | |
| *** zaitcev has joined #openstack-dev | 17:02 | |
| *** maplebed has joined #openstack-dev | 17:03 | |
| *** ewindisch has quit IRC | 17:03 | |
| openstackgerrit | A change to openstack/nova has been rejected: Add simple affinity filters, bro Update the filter so that context is part of the filter_properties, bro. Remove the Bro, get instance_id out of filter_properties['request_spec']['instance_properties'] Fixed whitespace. https://review.openstack.org/3215 | 17:06 |
| *** mikeyp has left #openstack-dev | 17:06 | |
| jaypipes | Vek: would you mind looking into this bug and providing your opinion on it? https://bugs.launchpad.net/glance/+bug/919240 | 17:07 |
| uvirtbot | Launchpad bug 919240 in glance "Glance API returns 401 response when disabled user's token is provided" [Undecided,New] | 17:07 |
| _0x44 | comstud: replied | 17:08 |
| openstackgerrit | A change to openstack/nova has been rejected: Add simple affinity filters, bro Update the filter so that context is part of the filter_properties, bro. Remove the Bro, get instance_id out of filter_properties['request_spec']['instance_properties'] Fixed whitespace. https://review.openstack.org/3215 | 17:08 |
| comstud | _0x44: replied twice | 17:09 |
| *** ewindisch has joined #openstack-dev | 17:10 | |
| *** deshantm_ has joined #openstack-dev | 17:10 | |
| _0x44 | comstud: replied now | 17:11 |
| bcwaldon | so meta | 17:12 |
| *** deshantm has quit IRC | 17:12 | |
| comstud | _0x44: Ok. I think instance_id goes in filter_properties vs instance_properties then | 17:12 |
| *** heckj has joined #openstack-dev | 17:13 | |
| _0x44 | comstud: It doesn't, it's in the instance_properties of the request_spec of the filter_properties. | 17:13 |
| comstud | _0x44: It doesn't exist right now, so it's not anywhere | 17:13 |
| comstud | _0x44: is this the instance id of the instanc ebeing scheduled? | 17:13 |
| comstud | _0x44: Or *another* instance? | 17:13 |
| _0x44 | No, it's the instance_id of another instance | 17:13 |
| comstud | _0x44: Then it definitely doesn't belong in instance_properties | 17:14 |
| comstud | that's the dict for the new instance | 17:14 |
| comstud | I'd add it to filter_properties | 17:14 |
| comstud | and I'd use 'uuid' if you're going to pass an id | 17:14 |
| comstud | :) | 17:14 |
| *** rmkdroid has quit IRC | 17:14 | |
| comstud | but i can live with 'id' for now | 17:14 |
| _0x44 | Damn, that's where it was in patchset1&2... but vishy suggested it get moved to the request_spec | 17:15 |
| comstud | request_spec should be things tied to the scheduled instance, really | 17:15 |
| _0x44 | What's in filter_properties['request_spec']? | 17:15 |
| _0x44 | Ok | 17:15 |
| comstud | the instance_type for new instance | 17:15 |
| comstud | the new instance's model (instance_properties) | 17:15 |
| comstud | (although the model is a dict at this point) | 17:16 |
| comstud | the number of instances to build (which needs to be removed) | 17:16 |
| comstud | and things like that | 17:16 |
| comstud | hmm | 17:16 |
| comstud | I guess I can see it being okay to put in request_spec | 17:16 |
| comstud | because we put things like 'original_host' in there | 17:16 |
| comstud | for resize | 17:17 |
| _0x44 | I'll just put it into filter-properties for now | 17:17 |
| comstud | but I'd name it something other than just "instance_id" | 17:17 |
| comstud | make it clear it's for another instance | 17:17 |
| _0x44 | "+ instance_uuid = filter_properties['affinity_instance_uuid']" | 17:17 |
| comstud | that works! | 17:17 |
| *** ewindisch has quit IRC | 17:17 | |
| *** Tribaal has quit IRC | 17:17 | |
| comstud | maybe we should add 'filter_properties' to the cast to the sheduler so the compute manager can populate some of it first before the scheduler does | 17:18 |
| comstud | i think that might make some sense | 17:19 |
| _0x44 | I think that was part of what vishy was planning bcwaldon to do | 17:19 |
| *** shevek_ has quit IRC | 17:19 | |
| _0x44 | But I only recall vishy telling bcwaldon to do everything | 17:19 |
| comstud | good man | 17:19 |
| comstud | i could probably prop that real quick right now tho | 17:20 |
| comstud | to help out with this filter stuff | 17:20 |
| bcwaldon | _0x44: it really was on my plate, but it doesn't bother me to let someone else do it | 17:20 |
| comstud | good, another thing I get to do for bcwaldon. | 17:20 |
| jaypipes | _0x44: nice work bro. | 17:20 |
| _0x44 | jaypipes: Thanks bro :P | 17:21 |
| jaypipes | pronounced "brah" | 17:21 |
| comstud | i was going to suggest using 'brah' :) | 17:21 |
| jaypipes | :) | 17:21 |
| jaypipes | comstud: us old fuddy-duddies may not have "gotten it" :) | 17:21 |
| comstud | haha | 17:21 |
| comstud | true | 17:21 |
| _0x44 | jaypipes: Why were you in jk0's closet? | 17:21 |
| jaypipes | lol | 17:21 |
| bcwaldon | broseph, guys | 17:22 |
| *** mdomsch has quit IRC | 17:22 | |
| jaypipes | comstud, bcwaldon: hey, want your opinion on these bugs: https://bugs.launchpad.net/glance/+bug/919257 and https://bugs.launchpad.net/glance/+bug/919240 | 17:22 |
| uvirtbot | Launchpad bug 919257 in glance "Glance Add Image API returns 400 response if storage is inaccessible" [Undecided,New] | 17:22 |
| jaypipes | comstud, bcwaldon: not sure if I buy the 500 server return... | 17:23 |
| comstud | ? | 17:23 |
| openstackgerrit | Christopher MacGown proposed a change to openstack/nova: Add simple affinity filters, bro Update the filter so that context is part of the filter_properties, bro. Remove the Bro, get instance_id out of filter_properties['request_spec']['instance_properties'] Fixed whitespace. https://review.openstack.org/3215 | 17:23 |
| jaypipes | comstud: do you think a 500 is the correct code to return there? | 17:23 |
| comstud | oh, missed the links | 17:23 |
| comstud | looking | 17:23 |
| jaypipes | comstud: maybe a 413 Request Entity Too Large? | 17:24 |
| jaypipes | for the first... | 17:24 |
| comstud | i disagree with a 500 on that first one | 17:24 |
| comstud | for sure | 17:24 |
| bcwaldon | jaypipes: why doesn't 401 make sense for the first bug? | 17:25 |
| jaypipes | comstud: when I see 500, I instinctively think the server is entirely kaput. | 17:25 |
| jaypipes | bcwaldon: not sure about that either... :) that's why I'm asking. | 17:25 |
| comstud | jaypipes: exactly re: 500 | 17:26 |
| jaypipes | bcwaldon: and I remember asking Vek about 401 vs 403 when he did that code... just can't remember what his response was | 17:26 |
| jaypipes | comstud: would you agree with a 413? | 17:26 |
| comstud | jaypipes: I'm not sure a '400' is correct response, either, though. 'Bad Request' | 17:26 |
| *** rnirmal has quit IRC | 17:26 | |
| _0x44 | jaypipes: I think the first one is really two bugs. If the storage is unavailable, it should be possibly 500? | 17:26 |
| jaypipes | _0x44: yeah, but a small image request may succeed, which is why I think a 413 is better... | 17:26 |
| jaypipes | _0x44: I mean, if I try and upload a 200G image to a server with 100G left for storage, a 500 would seem wrong... | 17:27 |
| comstud | if anything, it should be a 503 not a 500 | 17:27 |
| _0x44 | jaypipes: That's three bugs then? Storage _unavailable_ -> Oh shit what happened? Storage full -> 413. Permissions wrong -> 500 | 17:27 |
| comstud | if you're going with a 5xx error code | 17:27 |
| jaypipes | comstud: right, but the service itself is not necessarily unavailable... you could still GET or even POST with a smaller image.. | 17:28 |
| comstud | yeah | 17:28 |
| jaypipes | _0x44: dealing with this one right now: https://bugs.launchpad.net/glance/+bug/919257 | 17:28 |
| uvirtbot | Launchpad bug 919257 in glance "Glance Add Image API returns 400 response if storage is inaccessible" [Undecided,New] | 17:29 |
| comstud | that's why I don't like 5xx | 17:29 |
| * jaypipes wants a way to distinguish between "you can't use this server for anything" and "this server is full" | 17:29 | |
| _0x44 | jaypipes: He's documenting two error modes there. | 17:30 |
| _0x44 | jaypipes: Three if you're including no space left on the device. | 17:30 |
| _0x44 | comstud: I uploaded patchset 4 to that, or should I resubmit? | 17:31 |
| jaypipes | _0x44: sorry, wrong bug link... https://bugs.launchpad.net/glance/+bug/919255 | 17:31 |
| uvirtbot | Launchpad bug 919255 in glance "Glance Add Image API returns 400 response for disk full scenario" [Medium,Triaged] | 17:31 |
| jaypipes | comstud: ^^ | 17:31 |
| comstud | _0x44: I'll look... | 17:32 |
| _0x44 | jaypipes: Do you really want to expose the storage limit to the user? | 17:32 |
| jaypipes | _0x44: not sure... | 17:33 |
| jaypipes | _0x44: this is a bit of a group brainstorming session :) | 17:33 |
| *** andrewbogott has joined #openstack-dev | 17:35 | |
| *** andrewbogott has joined #openstack-dev | 17:35 | |
| *** reed has joined #openstack-dev | 17:35 | |
| comstud | jaypipes: Well, i'm with you in that I view a 5xx as a "shit, the server is broken" situation | 17:36 |
| comstud | store not available *might* fall into that, but not the others | 17:36 |
| smoser | jaypipes, ok... | 17:37 |
| smoser | hm... hld on a minute | 17:38 |
| comstud | maybe '412' for store not available | 17:38 |
| comstud | i dunno | 17:38 |
| smoser | i'm waiting for utlemming to join | 17:39 |
| jaypipes | comstud: meh, but 412 is really about a *client* precondition has failed, not a server condition.. | 17:39 |
| jaypipes | smoser: utlemming? | 17:40 |
| _0x44 | jaypipes: Add a client precondition "X-SHIT-WORKS" | 17:40 |
| *** eglynn_ has quit IRC | 17:40 | |
| Daviey | comstud: 5XX doesn't always mean the server is broken. | 17:40 |
| comstud | client sent the request under the assumption the store was available, but it's not? | 17:40 |
| comstud | :) | 17:40 |
| *** eglynn_ has joined #openstack-dev | 17:40 | |
| *** deshantm_ is now known as deshantm | 17:40 | |
| comstud | but yeah, i know | 17:40 |
| jaypipes | _0x44: lmao | 17:40 |
| smoser | jaypipes, utlemming is the other "ubuntu cloud images" guy | 17:40 |
| smoser | ie, 'smoser2' | 17:41 |
| smoser | :) | 17:41 |
| jaypipes | :) | 17:41 |
| smoser | or 2.0 | 17:41 |
| comstud | Daviey: Yeah, it's probably just a mindset issue for me... which I'm willing to change | 17:41 |
| jaypipes | Daviey: I was just remarking that a 500 specifically is something I've always considered "shit is fucked" | 17:41 |
| comstud | store not available could be 'shit is fucked' | 17:42 |
| jaypipes | probably a leftover from my days of having to deal with IIS... | 17:42 |
| jaypipes | comstud: indeed | 17:42 |
| openstackgerrit | Verification of a change to openstack/nova failed: Remove deprecated extension code. https://review.openstack.org/3190 | 17:42 |
| Daviey | jaypipes: 4XX = i've done something wrong, 5XX they are broken, 2XX is full of win. | 17:43 |
| Daviey | :) | 17:43 |
| jaypipes | :) | 17:43 |
| kbringard | lol, shit is fucked | 17:44 |
| kbringard | that's the technical term? | 17:44 |
| kbringard | :-p | 17:44 |
| *** betodalas_ has quit IRC | 17:45 | |
| Vek | anyone looking into a devstack problem? | 17:45 |
| *** bencherian_ has joined #openstack-dev | 17:45 | |
| Daviey | Vek: many people are looking into the problem that is devstack. | 17:46 |
| Vek | How about the specific problem that "git fetch" fails? :P | 17:46 |
| Kiall | i think Vek is talking about the devstack gate, and not devstack itself | 17:46 |
| comstud | Vek: that's the one... appears transient | 17:46 |
| comstud | it's gerrit | 17:47 |
| Vek | so, retry? | 17:47 |
| comstud | or jenkins | 17:47 |
| comstud | yea | 17:47 |
| *** andrea_ has quit IRC | 17:47 | |
| comstud | i had to retry a couple times last night | 17:47 |
| Vek | 'k. | 17:48 |
| *** Tribaal has joined #openstack-dev | 17:48 | |
| Vek | jaypipes: got a second? Need a little more detail on 919244 | 17:48 |
| Vek | specifically, is the error coming from api or registry? | 17:48 |
| *** utlemming has joined #openstack-dev | 17:49 | |
| jaypipes | Vek: not sure :) that's why I tagged that one with verification-needed ;) | 17:49 |
| jaypipes | Vek: I imagine adding a test case with the invalid request body included in the bug report will lead to the predicated fail... | 17:50 |
| smoser | ok. | 17:50 |
| smoser | jaypipes, now.. utlemming (aka smoser 2.0) arrives | 17:51 |
| utlemming | lol | 17:51 |
| smoser | so.. i was talking to utlemming, and if the only issue (it was my primary issue) with the public glance api server | 17:51 |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/python-keystoneclient: Support for version discovery https://review.openstack.org/2641 | 17:51 |
| Vek | jaypipes: yeah, but the registry does the json deserialize after verification of authorization | 17:51 |
| smoser | was that it would not cache locally, he and i considered that we could just use a transparent (or configured on the nodes) squid proxy | 17:51 |
| Vek | I suspect the problem is in the API | 17:52 |
| *** rajaram has quit IRC | 17:52 | |
| *** rajaram1 has joined #openstack-dev | 17:52 | |
| smoser | so that each node's request to glance.ubuntu.com/images (or what ever) would get proxied and downloaded and locally cached that way. | 17:52 |
| jaypipes | smoser: right. | 17:52 |
| smoser | does that seem like a reasonable solution? | 17:52 |
| jaypipes | smoser: yes. isn't that what we discussed earlier on the phone? | 17:52 |
| Kiall | humm, does anyone know if there is a bug in the diablo block migration code that prevents the libvirt nwfilter file from being moved with the instance? | 17:53 |
| * Vek ponders further | 17:53 | |
| Kiall | (and deleted if its copied over manually) | 17:53 |
| smoser | the only issue that i saw as that we'd need for the nodes to be able to choose a proxy based on url i think (or some how mitigate the likelyhood that there are 2 glance servers) | 17:53 |
| smoser | jaypipes, i must have missed something then. | 17:53 |
| smoser | i'm just saying, that a simple http_proxy variable would need to be added to the nova nodes so they would proxy image downloads through something centralized. | 17:54 |
| smoser | but possibly a *simple* value wouldn't be enough... they might have to know "oh, for things to that local glance server, i should get them directly, but for intarweb stuff, i need to proxy" (to avoid a local proxy of a local resource... which might be able to be resolved elsewhere). | 17:55 |
| *** derekh has quit IRC | 17:55 | |
| jaypipes | smoser: but I thought that we agreed that the nova compute nodes would *only* point to local glance nodes, and that we'd write a "uec-get-base-images" script that would pull images from glance.ubuntu.com/images and populate local glance nodes with copies of those images? | 17:55 |
| smoser | jaypipes, right. thats what you and i came to. | 17:56 |
| jaypipes | smoser: and the squid servers would be in front of glance.ubuntu.com | 17:56 |
| smoser | ah... see i'm not explaining myself right. | 17:56 |
| jaypipes | sorry :( I'm a bit thick. | 17:56 |
| smoser | the "local cloud provider" (ie, smoser's basement) would just configure th enova nodes to do their image downloads via 'smoser-local-squidproxy:3128' | 17:57 |
| smoser | and then we could do the full glance server thing | 17:57 |
| smoser | but the download from glance.ubutnu.com would only happen once | 17:58 |
| smoser | because squid would be caching it locally | 17:58 |
| smoser | where locally means "smoser's basement" not "the node" | 17:58 |
| jaypipes | I see | 17:58 |
| smoser | that seemed to me to transparently solve the "I want to avoid external bandwidth usage" issue | 17:59 |
| jaypipes | but why wouldn't we just modify the "uec-get-base-images" script to point to smoser-local-squidproxy:3128 instead of glance.ubuntu.com? | 17:59 |
| smoser | ah. | 17:59 |
| jaypipes | smoser: not modify the script, but add a CLI option for the source ... | 17:59 |
| jaypipes | you know what I mean... | 17:59 |
| smoser | so we could do that, yes. | 17:59 |
| smoser | i was thinking that with "uec-get-base-images", you'd just pull the images (from cloud-images.ubuntu.com), populate local glance, and never again need cloud-images.ubuntu.com | 18:00 |
| smoser | *nor* would we actually need a glance server | 18:00 |
| smoser | on glance.ubuntu.com | 18:00 |
| openstackgerrit | A change to openstack/nova has been rejected: Add simple affinity filters, bro. Update the filter so that context is part of the filter_properties, bro. Changed the instance_id -> affinity_instance_uuid so it's clear https://review.openstack.org/3215 | 18:01 |
| jaypipes | smoser: I see. | 18:01 |
| smoser | as we have reasonable data at http://cloud-images.ubuntu.com/query | 18:01 |
| smoser | ie, that'd not be a glance client, just an http client. | 18:01 |
| smoser | but i see how you mistook tat. | 18:01 |
| jaypipes | sure | 18:01 |
| utlemming | where I think having the glance server is beneficial is that it means that a cloud can consume the images as needed, instead of having to syncronize every so often | 18:01 |
| *** adjohn has joined #openstack-dev | 18:01 | |
| jaypipes | utlemming: yes | 18:01 |
| smoser | but the glance.ubutnu.com has nicities | 18:01 |
| kbringard | just setup torrents and write a handler for glance to connect to the swarm and pull down the images the user specifies in their script ;-) | 18:01 |
| smoser | nice-ities ? | 18:01 |
| smoser | anyway. | 18:01 |
| *** ewindisch has joined #openstack-dev | 18:01 | |
| jaypipes | kbringard: yeah, yeah, we discussed that too :) | 18:02 |
| utlemming | my concern is that when we publish new images (i.e. new release or spin new images), then there will be a delta between cloud providers | 18:02 |
| kbringard | sorry, late to the party | 18:02 |
| jaypipes | kbringard: but first step is getting UEC base images into the local glance nodes as efficiently as possible :) | 18:02 |
| utlemming | it would be nice to enable cloud-providers the ability to do nothing to get the latest and greatest | 18:02 |
| smoser | right. and the glance server, with multi-glance-server function in horizon would immediately *show* the images, although the first would be slow to boot. | 18:02 |
| *** bencherian_ has quit IRC | 18:03 | |
| smoser | the one thing i fear in the glance.ubuntu.com server is its reliability | 18:03 |
| smoser | if we go down, then possibly your horizon doesn't work that well, and/or you can't boot a new image. | 18:03 |
| jaypipes | ya | 18:03 |
| utlemming | and slow to boot for the first is better than waiting for a cron job or worse, waiting for a manual sync | 18:03 |
| _0x44 | comstud: host is an ip address, and so is h2. We build a cidr network that contains host using the cidr suffix. If h2 is in that network it returns true. | 18:03 |
| smoser | and the full "mirror it local" removes that. | 18:03 |
| smoser | so anyway... | 18:04 |
| comstud | _0x44: Huh. 'host' is the compute host the instance is on | 18:04 |
| smoser | i'll try to write this all down more clearly and send a mail | 18:04 |
| comstud | _0x44: Are you naming your compute hosts by IP address? | 18:04 |
| jaypipes | smoser: yeah, but "mirror it local" has the issue utlemming brought up about cache staleness when a new base image is pushed. | 18:04 |
| comstud | _0x44: Seems like a strange hostname | 18:04 |
| smoser | but i wanted to get your feeling as to how absurd/reasonable it would be to have nova-compute instances proxied downloads. | 18:04 |
| smoser | jaypipes, absolutely. | 18:04 |
| kbringard | have each glance server be a seeder… so the user specifies a torrent to import into one of their glance servers, at which point all the servers in the glance clusters, wherever they may be start distributing the image across the private "client" torrent network | 18:04 |
| _0x44 | comstud: possibly... | 18:04 |
| smoser | there are benefits to both. | 18:05 |
| _0x44 | comstud: Can make that do a db request there, but we don't have to do a cidr lookup on h2. | 18:05 |
| utlemming | jaypipes: I'm not sure I follow -- each image is named uniquely, which would prevent stale images | 18:05 |
| _0x44 | Just get two ip addresses | 18:05 |
| jaypipes | kbringard: yes, that is how antonym and s1rp did it inside RAX IIRC.. | 18:05 |
| *** eglynn__ has joined #openstack-dev | 18:05 | |
| jaypipes | utlemming: an image is not named uniquely, no. | 18:05 |
| smoser | utlemming, s/stale images/stale data/ | 18:05 |
| jaypipes | utlemming: it has a unique ID | 18:05 |
| smoser | our images are named uniquely | 18:05 |
| comstud | 0x44: I guess I just need more information on what you're trying to accomplish. You're trying to look at the IP address assigned to the instance, no? | 18:05 |
| _0x44 | comstud: NO. | 18:05 |
| smoser | we'd populate them with a well defined (and used on EC2) naming convention | 18:06 |
| comstud | 0x44: What's the IP address of? compute host? | 18:06 |
| _0x44 | comstud: It's trying to find a host that's in the same cidr range as the compute host of your affinity instance | 18:06 |
| jaypipes | smoser: you have to remember that once an image is in glance, it is immutable. | 18:06 |
| jaypipes | smoser: the only thing you can do is the following: | 18:06 |
| smoser | thats fine. | 18:06 |
| smoser | we consider images immutable on ec2 | 18:07 |
| jaypipes | smoser: a) "tag" an image in glance.ubuntu.com with a custom property like "latest 11.10" or similar | 18:07 |
| smoser | i never re-publish something under the same name | 18:07 |
| comstud | 0x44: ah ok. then I guess if you have FLAGS.host on your compute workers set to an IP address, this would work. | 18:07 |
| comstud | 0x44: Seems like a strange assumption for the filter, though | 18:07 |
| jaypipes | smoser: b) have the "uec-get-base-images" script poll the glance.ubuntu.com registry for the image IDs for images tagged with the "latest XXX" properties | 18:07 |
| *** eglynn_ has quit IRC | 18:07 | |
| openstackgerrit | Verification of a change to openstack/nova failed: Remove deprecated extension code. https://review.openstack.org/3190 | 18:08 |
| _0x44 | comstud: It's _simple_... | 18:08 |
| smoser | jaypipes, yeah.. | 18:08 |
| jaypipes | smoser: have that script compare some local cache of those base image IDs, and if anything changed, pull the "updated" image | 18:08 |
| _0x44 | comstud: It's a reasonable assumption that a host on the same /28 are close to each other. | 18:08 |
| smoser | so one benefit of the glance.ubutnu.com is that you can then write a glance client that determines which image-id you need to launch. | 18:08 |
| _0x44 | comstud: at least for a simple affinity scheduler | 18:08 |
| comstud | 0x44: i don't argue the value | 18:09 |
| smoser | wel.. | 18:09 |
| smoser | anyway. | 18:09 |
| jaypipes | smoser: I suppose you could also do that with the name of the image, like utlemming suggested, but there isn't any uniqueness enforced in the DB for name (in fact, Name can be None!) | 18:09 |
| comstud | 0x44: I'm just saying this only works if you've set FLAGS.host on your compute instances to an IP address. | 18:09 |
| comstud | 0x44: Which seems like a strange assumption | 18:09 |
| comstud | 0x44: Since the default in nova for FLAGS.host is the hostname of the host... which I doubt is an IP address for 99.9% of people | 18:09 |
| smoser | well, i do like the tag, it makes things easier, and i like that glance.ubuntu.com makes that easy to get a global unique glance id that you can then say "launch this" | 18:09 |
| smoser | which means the smae thing on my cloud and yours | 18:10 |
| smoser | (since they both know aobut glance.ubuntu.com) | 18:10 |
| _0x44 | comstud: So are you suggesting a DNS lookup? | 18:11 |
| comstud | 0x44: In which case, I'd say the filter doesn't belong in nova and should be packaged by whomever needs it. I think there must be a better way to do this generically. | 18:11 |
| comstud | 0x44: I'd love to avoid that, too! ;) | 18:11 |
| comstud | Hm. | 18:11 |
| *** Gordonz has joined #openstack-dev | 18:11 | |
| *** Gordonz has quit IRC | 18:12 | |
| *** jdurgin has joined #openstack-dev | 18:12 | |
| _0x44 | comstud: I think _an_ affinity filter better than "install on the same host" and "don't install on the same host" is useful for nova. | 18:13 |
| jaypipes | smoser: so, we good? got enough to move forward? | 18:13 |
| smoser | jaypipes, yeah, i think so. | 18:13 |
| smoser | i'll send a mail. | 18:13 |
| *** Gordonz has joined #openstack-dev | 18:13 | |
| smoser | so now i'm on the path of glance.ubutnuc.om | 18:13 |
| kbringard | smoser: it would also allow the glance-pruner to key off of something to know if an image has been updated | 18:13 |
| smoser | and ii'mk going to need someone to do some horizen work | 18:13 |
| smoser | for multiple glance servers in it. | 18:13 |
| comstud | 0x44: Perhaps we can brainstorm with vish.. I know that at some point here I'm going to be adding an 'affinity ID' which would solve this | 18:13 |
| _0x44 | comstud: This seems like the simplest way that doesn't require any assumptions about "nearness" | 18:13 |
| jaypipes | smoser: well, cool then :) | 18:13 |
| kbringard | so you don't end up with 20 iterations of the same image as it gets updated | 18:13 |
| *** mjfork has quit IRC | 18:13 | |
| jaypipes | smoser: I can do the Horizon work I think. | 18:14 |
| comstud | 0x44: affinity ID will be a unique ID which will map to a set of hosts (or really, in our case, a zone) | 18:14 |
| _0x44 | comstud: That requires that you expose hosts to the user. | 18:14 |
| comstud | 0x44: If we have that, then this check is extremely simple | 18:14 |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/python-keystoneclient: Support for version and extension discovery https://review.openstack.org/2641 | 18:15 |
| comstud | How is that any different than what you have here? | 18:15 |
| comstud | it's just comparing a unique ID vs a CIDR | 18:15 |
| *** zns has quit IRC | 18:17 | |
| *** pixelbeat has quit IRC | 18:17 | |
| _0x44 | comstud: This isn't intended to be the end-all affinity filter. | 18:17 |
| _0x44 | comstud: It's a stupid simple one that takes two instances and compares some value of their hosts that just happens to be ip addresses. | 18:17 |
| comstud | yeah, in your specific case | 18:17 |
| comstud | which is probably not the norm | 18:17 |
| comstud | in which case, it feels like it doesn't belong in nova. it's something specific to your environment. | 18:18 |
| _0x44 | comstud: This isn't for my use-case | 18:18 |
| *** mjfork has joined #openstack-dev | 18:19 | |
| comstud | If it's vish's, I can discuss with him later | 18:19 |
| _0x44 | comstud: I haven't implemented anything in nova that is specific to Piston Cloud. | 18:19 |
| _0x44 | comstud: Actually, _everything_ I've written for OpenStack is wholly separate from Piston Cloud's needs. | 18:19 |
| _0x44 | The use-case for that filter was, "There needs to be something more complex than the first two but not so complex that people try to use it" | 18:20 |
| comstud | Sure, i have no worries there | 18:20 |
| comstud | no need to get defensive ;) | 18:20 |
| _0x44 | No worries there, but you still assume that I'm trying to sneak code in that is only useful to my company? | 18:20 |
| comstud | i'm just saying that this filter would only work under a very specific configuration of nova that I suspect only .001% of the people using it would do | 18:21 |
| _0x44 | Unless I change it so that it does a DNS lookup for things that aren't IPs, you mean. | 18:21 |
| comstud | so I'm trying to figure out something that satisifies the use case for more than that percentage | 18:21 |
| comstud | :) | 18:21 |
| _0x44 | Which I've already suggested doing because it keeps the simplicity. | 18:21 |
| _0x44 | And I can change the name to SimpleExpensiveAffinityFilter or something | 18:21 |
| comstud | i mean, a DNS query would work, as long as you have resolv.conf set up properly and you entered the hosts names in DNS | 18:22 |
| comstud | resolv.conf would need search domain | 18:22 |
| comstud | you could also call to the compute manager and ask it for its IP address | 18:22 |
| comstud | which it could probably figure out more easily | 18:22 |
| _0x44 | That makes sense, I'll do that. | 18:23 |
| comstud | i'm not sure i like either of those, but i guess they'd satisfy my concerns | 18:23 |
| comstud | even if it slows the filter. | 18:23 |
| comstud | then we can fix it all when there is a larger affinity solution | 18:24 |
| comstud | which is on our list | 18:24 |
| _0x44 | Right, that was the idea. | 18:24 |
| comstud | yeah, i get it. i'd have a different opinion about the current patch if FLAGS.host defaulted to the host's IP address. | 18:25 |
| comstud | because then it would work out of the box | 18:26 |
| openstackgerrit | Verification of a change to openstack/glance failed: Swallow exception on unsupported image deletion. https://review.openstack.org/3192 | 18:27 |
| comstud | i realize i'm being picky, but I have opinions about what should and shouldn't be in nova. I don't think we want 1000 filters that only a small select group use... when it can be a filter packaged separately outside of nova. | 18:28 |
| *** _adjohn has joined #openstack-dev | 18:32 | |
| *** _adjohn has quit IRC | 18:33 | |
| *** adjohn has quit IRC | 18:33 | |
| *** adjohn has joined #openstack-dev | 18:33 | |
| *** rajaram1 has quit IRC | 18:34 | |
| *** pixelbeat has joined #openstack-dev | 18:35 | |
| openstackgerrit | Dean Troyer proposed a change to openstack/nova: Fix nova-manage floating list (fixes bug 919288) https://review.openstack.org/3227 | 18:35 |
| uvirtbot | Launchpad bug 919288 in devstack "nova-manage floating list fails" [Undecided,New] https://launchpad.net/bugs/919288 | 18:36 |
| *** bencherian has joined #openstack-dev | 18:37 | |
| *** stuntmachine has joined #openstack-dev | 18:38 | |
| *** zns has joined #openstack-dev | 18:39 | |
| *** Ryan_Lane has joined #openstack-dev | 18:39 | |
| *** kyriakos has left #openstack-dev | 18:40 | |
| *** CarlosM has left #openstack-dev | 18:44 | |
| *** rbasak has quit IRC | 18:47 | |
| openstackjenkins | Project nova-docs build #1988: SUCCESS in 3 min 23 sec: https://jenkins.openstack.org/job/nova-docs/1988/ | 18:49 |
| openstackjenkins | naveedm9: Validating image id for rebuild. | 18:49 |
| *** novas0x2a|laptop has joined #openstack-dev | 18:50 | |
| ayoung | Do we have a good example of where we have to wrap a blocking IO call to make it work with Eventlet? | 18:57 |
| *** stuntmachine has quit IRC | 18:58 | |
| *** stuntmachine has joined #openstack-dev | 18:59 | |
| *** darraghb has quit IRC | 19:03 | |
| openstackjenkins | Project nova-docs build #1989: SUCCESS in 3 min 32 sec: https://jenkins.openstack.org/job/nova-docs/1989/ | 19:04 |
| openstackjenkins | * Chuck Short: Drop FK constraint if it exists in migration 064 | 19:04 |
| openstackjenkins | * Vishvananda Ishaya: Fix missing imports and bad call caught by pyflakes | 19:04 |
| Kiall | am I right in thinking diablo block migration does not work correctly when using keystone? | 19:06 |
| *** lloydde has joined #openstack-dev | 19:06 | |
| *** pixelbeat has quit IRC | 19:08 | |
| *** rkukura has quit IRC | 19:16 | |
| openstackjenkins | Project nova-docs build #1990: SUCCESS in 3 min 27 sec: https://jenkins.openstack.org/job/nova-docs/1990/ | 19:19 |
| openstackjenkins | jason: Allow Quantum Manager to run in "Flat" mode. | 19:19 |
| *** andrewsben has joined #openstack-dev | 19:19 | |
| *** jog0 has joined #openstack-dev | 19:25 | |
| *** mdomsch_ has joined #openstack-dev | 19:25 | |
| openstackgerrit | Kevin L. Mitchell proposed a change to openstack/glance: Return 400 if registry returns 400. https://review.openstack.org/3228 | 19:32 |
| Vek | jaypipes: that should fix the problem | 19:32 |
| jaypipes | Vek: cheerio | 19:32 |
| *** crobinso has joined #openstack-dev | 19:33 | |
| openstackjenkins | Project nova-docs build #1991: SUCCESS in 3 min 43 sec: https://jenkins.openstack.org/job/nova-docs/1991/ | 19:34 |
| openstackjenkins | Isaku Yamahata: nova/network: pass network_uuid to linuxnet_interface_driver and vif driver | 19:34 |
| Vek | can I get mtaylor or jeblair to look at https://review.openstack.org/#change,3190 please? devstack keeps failing to git fetch | 19:34 |
| openstackgerrit | Verification of a change to openstack/nova failed: Remove deprecated extension code. https://review.openstack.org/3190 | 19:37 |
| Vek | OK, this time it fetched, but claims nova-api failed to start. A failure I cannot replicate. And of course the error messages are singularly unhelpful. | 19:39 |
| openstackgerrit | Dean Troyer proposed a change to openstack/nova: Fix nova-manage floating list (fixes bug 919288) https://review.openstack.org/3227 | 19:41 |
| uvirtbot | Launchpad bug 919288 in nova "nova-manage floating list fails (dup-of: 918804)" [Undecided,In progress] https://launchpad.net/bugs/919288 | 19:41 |
| uvirtbot | Launchpad bug 918804 in nova "'nova-manage floating list' calls undefined method 'floating_ip_get_all'" [Medium,In progress] https://launchpad.net/bugs/918804 | 19:41 |
| openstackgerrit | Dean Troyer proposed a change to openstack/nova: Fix nova-manage floating list (fixes bug 919288) https://review.openstack.org/3227 | 19:42 |
| uvirtbot | Launchpad bug 919288 in nova "nova-manage floating list fails (dup-of: 918804)" [Undecided,In progress] https://launchpad.net/bugs/919288 | 19:42 |
| uvirtbot | Launchpad bug 918804 in nova "'nova-manage floating list' calls undefined method 'floating_ip_get_all'" [Medium,In progress] https://launchpad.net/bugs/918804 | 19:42 |
| openstackgerrit | Zhongyue Luo proposed a change to openstack/nova: nova flags list tool https://review.openstack.org/3210 | 19:44 |
| *** bencherian has quit IRC | 19:48 | |
| jaypipes | Vek: you're a stubborn mule aren't ya. ;P | 19:49 |
| * Vek plants his cleats | 19:51 | |
| jk0 | lol | 19:52 |
| *** andrewsben has quit IRC | 19:53 | |
| *** adjohn has quit IRC | 19:53 | |
| jaypipes | :) | 19:59 |
| *** zzed has quit IRC | 20:00 | |
| *** rkukura has joined #openstack-dev | 20:00 | |
| *** lloydde has quit IRC | 20:04 | |
| *** lloydde has joined #openstack-dev | 20:05 | |
| openstackgerrit | David Kranz proposed a change to openstack/tempest: Don't set multiple images if image_ref_alt is the same as image_ref. Fix typo in skip_unless_addr. https://review.openstack.org/3183 | 20:07 |
| *** sleepsonthefloo has joined #openstack-dev | 20:07 | |
| adam_g | looking at the glance-{registry, api}-paste.ini files for the first time... is the idea that you just specify the corresponding flavor in glance-{registry, api}.conf and not need to worry about messing around with pipeline configuration in the .ini files anymore? | 20:08 |
| adam_g | jaypipes: ^ ? | 20:08 |
| *** stuntmachine has quit IRC | 20:08 | |
| jaypipes | adam_g: yep. | 20:09 |
| openstackgerrit | David Kranz proposed a change to openstack/tempest: Don't set multiple images if image_ref_alt is the same as image_ref. Fix typo in skip_unless_addr. https://review.openstack.org/3183 | 20:10 |
| jaypipes | adam_g: eglynn__ is a good person to talk with about that new functionality | 20:10 |
| adam_g | jaypipes: ah, cool | 20:10 |
| openstackgerrit | Zhongyue Luo proposed a change to openstack/nova: nova flags list tool https://review.openstack.org/3210 | 20:10 |
| eglynn__ | adam_g: new glance functionality, or? | 20:10 |
| eglynn__ | yep that would be jay | 20:10 |
| smoser | anyone have a link to how to use "config drive" ? | 20:11 |
| eglynn__ | a-ha, I see you mean the paste config split? | 20:11 |
| eglynn__ | that would be me ... | 20:11 |
| adam_g | im wondering (from a packaging POV), if it'd make sense to have a seperate place to configure keystone details/credentials.. then, the api paste flavors can be updated with package upgrades based on whats is currently supported | 20:12 |
| adam_g | eglynn__: yeah, we're updating ubuntu packaging ATM and just ran into the split for the first time | 20:12 |
| dprince | adam_g: I like the idea of that.... | 20:12 |
| eglynn__ | yeah that was the idea, to avoid the need for users to directly edit the paste config in order to select a particular pipeline | 20:13 |
| adam_g | i like it | 20:13 |
| eglynn__ | (and be aware of ordering, what filters map on to whatever QoS etc.) | 20:13 |
| openstackgerrit | Zhongyue Luo proposed a change to openstack/nova: nova flags list tool https://review.openstack.org/3210 | 20:13 |
| dprince | eglynn: I'm a big fan of the Glance split. But taking it a step further and splitting out keystone creds might be even better. | 20:13 |
| adam_g | eglynn__: the issue is, if i've editted the keystone credentials in those files and packaging installs a new version of glance with updated pipelines, packaging will not write the updated config. | 20:14 |
| openstackgerrit | sleepsonthefloor proposed a change to openstack-dev/devstack: Downcase Admin user role to support nova's new policy engine https://review.openstack.org/3229 | 20:14 |
| dprince | eglynn/adam_g: This isn't just a glance thing though. We should do the same for nova as well. | 20:14 |
| eglynn__ | yeah, the auth token being embedded in the paste config is not nice | 20:14 |
| smoser | _0x44, ? how can i use config-drive ? is it exposed at all through nova-client ? | 20:14 |
| adam_g | dprince: is nova-api is the only consumer of the keystone credentials currently? | 20:15 |
| *** adjohn has joined #openstack-dev | 20:16 | |
| *** stuntmachine has joined #openstack-dev | 20:16 | |
| dprince | adam_g: nova-api is the only consumer of nova-api.ini. | 20:16 |
| eglynn__ | adam_g: by consumer of keystone credentials, do you mean client of keystone to verify auth tokens? | 20:16 |
| adam_g | eglynn__: yes. the only package that would potentially need the paste config installed | 20:17 |
| dprince | adam_g: which is where the keystone creds are we are talking about. Keystone auth tokens however get passed around to other services in nova. | 20:17 |
| eglynn__ | adam_g: glance is also a keystone client | 20:17 |
| adam_g | for glance i'd think there would need to be 2 files with keysotne information, one for glance-api and one for glance-registry. | 20:18 |
| eglynn__ | right | 20:18 |
| eglynn__ | (that's the current setup) | 20:18 |
| adam_g | which sorta sucks, but is better than having packaging unable to keep paste deploy pipelines up to date | 20:18 |
| dprince | adam_g/eglynn__: its also important to keep in mind Glance API and Glance Registry may not always run on the same node. A larger deployment will typically have many API servers for a single registry. | 20:20 |
| eglynn__ | the pipelines are not currently sufficiently common to share across glance-api and glance-registry | 20:20 |
| eglynn__ | (app endpoint differs) | 20:20 |
| eglynn__ | dprince: true that | 20:21 |
| adam_g | dprince: right | 20:21 |
| openstackgerrit | Anne Gentle proposed a change to openstack/openstack-manuals: Fixes build failure on epub build https://review.openstack.org/3230 | 20:21 |
| openstackgerrit | Dolph Mathews proposed a change to openstack/keystone: Added Vary header to support caching (bug 913895) https://review.openstack.org/3231 | 20:22 |
| uvirtbot | Launchpad bug 913895 in keystone "Return HTTP Vary header" [Wishlist,Confirmed] https://launchpad.net/bugs/913895 | 20:22 |
| *** kbringard has quit IRC | 20:23 | |
| *** kbringard has joined #openstack-dev | 20:24 | |
| *** LinuxJedi has quit IRC | 20:25 | |
| *** eglynn__ has quit IRC | 20:29 | |
| *** eglynn__ has joined #openstack-dev | 20:30 | |
| *** jakedahn has joined #openstack-dev | 20:30 | |
| *** gregburek has joined #openstack-dev | 20:32 | |
| _0x44 | smoser: It isn't exposed through nova client | 20:39 |
| smoser | so in order to use i have to write a nova client ? | 20:40 |
| openstackgerrit | Dan Prince proposed a change to openstack/openstack-chef: Updates to support glance paste changes. https://review.openstack.org/3233 | 20:41 |
| _0x44 | smoser: It's exposed through the API, novaclient wasn't part of nova when I wrote the code. | 20:41 |
| smoser | thats fine. do you have any example code, justso i can not re-write it? | 20:42 |
| *** dprince has quit IRC | 20:42 | |
| _0x44 | curl -i -X POST -d '{"server" : {…, "config_drive":"true"}}' host -H 'X-Auth-token: <t>' | 20:44 |
| _0x44 | curl -i -X POST -d '{"server" : {…, "config_drive":"image-ref"}}' … | 20:44 |
| *** lloydde has quit IRC | 20:45 | |
| openstackgerrit | sleepsonthefloor proposed a change to openstack-dev/devstack: Fix test that checks whether insances from previous runs need to be removed https://review.openstack.org/3234 | 20:46 |
| *** gyee has joined #openstack-dev | 20:49 | |
| *** dubsquared has quit IRC | 20:50 | |
| sleepsonthefloo | crobinso - hey anthony here. just want to say hello, and let you know that I'm working on openstackx servers calls today I think we should be able to have the dep removed by e3 | 20:52 |
| *** byeager has quit IRC | 20:54 | |
| crobinso | sleepsonthefloo: hi! yeah sounds good. gonna try and get the novaclient usage bits submitted then going to do another rev of the nova flavors patch | 20:54 |
| *** rnirmal has joined #openstack-dev | 20:55 | |
| sleepsonthefloo | crobinso - coolio - I'll keep an eye out for that branch and give a try at plugging it into horizon. | 20:57 |
| openstackgerrit | Emma Grace Steimann proposed a change to openstack/horizon: Added specific style for error box. * fixes bug 900432 https://review.openstack.org/3235 | 20:58 |
| uvirtbot | Launchpad bug 900432 in horizon "keypair name text box on create keypair screen goes into description text" [Low,Confirmed] https://launchpad.net/bugs/900432 | 20:58 |
| crobinso | sleepsonthefloo: unfortunately the usage bits in horizon seem to be a bit in flux so its hard to see it all at work. currently trying to view an individual tenants usage doesn't do much useful in horizon | 20:58 |
| crobinso | sleepsonthefloo: before or after my patch. but I assume it's on someones radar | 20:59 |
| openstackgerrit | Brian Waldon proposed a change to openstack/nova: Remove admin_only ext attr in favor of authz https://review.openstack.org/3236 | 20:59 |
| *** _adjohn has joined #openstack-dev | 20:59 | |
| *** adjohn has quit IRC | 21:02 | |
| *** _adjohn is now known as adjohn | 21:02 | |
| openstackgerrit | Kiall Mac Innes proposed a change to openstack/nova: Fixes bug #919390 - Block Migration fails when keystone is un use. https://review.openstack.org/3237 | 21:03 |
| uvirtbot | Launchpad bug 919390 in nova "Block migration fails when keystone is in use." [Undecided,New] https://launchpad.net/bugs/919390 | 21:03 |
| openstackgerrit | Chmouel Boudjnah proposed a change to openstack/swift: PEP8 fixes. https://review.openstack.org/3238 | 21:07 |
| openstackgerrit | Kiall Mac Innes proposed a change to openstack/nova: Fixes bug #919390 - Block Migration fails when keystone is un use. https://review.openstack.org/3237 | 21:12 |
| uvirtbot | Launchpad bug 919390 in nova "Block migration fails when keystone is in use." [Undecided,In progress] https://launchpad.net/bugs/919390 | 21:12 |
| openstackgerrit | Jay Pipes proposed a change to openstack/glance: Bug#911599 - Location field wiped on update https://review.openstack.org/2952 | 21:16 |
| openstackgerrit | Gabriel Hurley proposed a change to openstack/horizon: Auto-remove auto-generated .rst files for deleted .py files. https://review.openstack.org/3216 | 21:19 |
| openstackgerrit | Gabriel Hurley proposed a change to openstack/horizon: Added a contributing guide to the Horizon docs. https://review.openstack.org/3217 | 21:19 |
| *** markvoelker has quit IRC | 21:21 | |
| openstackgerrit | Gabriel Hurley proposed a change to openstack/horizon: Tooltip implementation cleanup. https://review.openstack.org/3218 | 21:22 |
| *** _adjohn has joined #openstack-dev | 21:23 | |
| openstackgerrit | Verification of a change to openstack/nova failed: Fix nova-manage floating list (fixes bug 918804) https://review.openstack.org/3227 | 21:23 |
| uvirtbot | Launchpad bug 918804 in nova "'nova-manage floating list' calls undefined method 'floating_ip_get_all'" [Medium,In progress] https://launchpad.net/bugs/918804 | 21:23 |
| *** _adjohn has quit IRC | 21:24 | |
| *** adjohn has quit IRC | 21:24 | |
| *** adjohn has joined #openstack-dev | 21:24 | |
| *** deshantm has quit IRC | 21:29 | |
| *** bencherian has joined #openstack-dev | 21:33 | |
| openstackgerrit | Verification of a change to openstack/nova failed: Support filter based on CPU core (over)allocation https://review.openstack.org/3209 | 21:33 |
| *** zul has quit IRC | 21:36 | |
| sleepsonthefloo | crobinso - yes i see that. there has been some refactoring in horizon that has munged some of the admin features. I was going to do a quick pass to try to bring some of it back, but them again, to some extent missing features make our job easier right now | 21:37 |
| *** david-kranz has joined #openstack-dev | 21:40 | |
| *** kbringard has quit IRC | 21:43 | |
| sleepsonthefloo | crobinso - one example of that is that the admin instance list no longer shows host (which is very useful for admins) - but I think I may bug that for e4 - it requires a nova extension | 21:43 |
| openstackgerrit | Jay Pipes proposed a change to openstack/glance: Bug #909574: Glance does not sanity-check given image size on upload https://review.openstack.org/2727 | 21:43 |
| uvirtbot | Launchpad bug 909574 in glance "Glance does not sanity-check given image size on upload" [Medium,In progress] https://launchpad.net/bugs/909574 | 21:43 |
| openstackgerrit | Verification of a change to openstack/glance failed: Bug#911599 - Location field wiped on update https://review.openstack.org/2952 | 21:43 |
| openstackgerrit | Jay Pipes proposed a change to openstack/tempest: Adds Images API tests https://review.openstack.org/2884 | 21:49 |
| *** bencherian has quit IRC | 21:51 | |
| *** _adjohn has joined #openstack-dev | 21:52 | |
| *** bencherian has joined #openstack-dev | 21:54 | |
| *** adjohn has quit IRC | 21:54 | |
| *** LinuxJedi has joined #openstack-dev | 21:55 | |
| *** eglynn has joined #openstack-dev | 21:55 | |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Create nova cert extension for x509 support https://review.openstack.org/3199 | 21:55 |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Adds extension for retrieving certificates https://review.openstack.org/3200 | 21:55 |
| *** bencherian has quit IRC | 21:56 | |
| *** _adjohn has quit IRC | 21:56 | |
| *** adjohn has joined #openstack-dev | 21:56 | |
| vishy | comstud: quick question if you got a sec | 21:57 |
| *** eglynn__ has quit IRC | 21:57 | |
| comstud | vishy: sure | 21:58 |
| *** jaypipes is now known as jaypipes-afk | 21:58 | |
| vishy | comstud: so the part of az-scheduler which allows an admin to specify a host | 21:58 |
| vishy | i realized should be separate | 21:59 |
| vishy | because if an admin specifies a host, it should just skip all of the filters entirely | 21:59 |
| *** stuntmachine has quit IRC | 21:59 | |
| vishy | so I'm thinking of doing something like ignored_host where we add a required_host to the request spec | 21:59 |
| vishy | and if it is set, it short circuits all the filtering and weighing and goes directly to that host | 22:00 |
| vishy | does that seem reasonable? ^^ | 22:00 |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Adds extension for retrieving certificates https://review.openstack.org/3200 | 22:00 |
| comstud | vishy: I think so. I'm in the middle of adding filter_properties as being passed to run_instance from compute manager | 22:00 |
| vishy | comstud: about _0x44's patch | 22:01 |
| comstud | In which case, I think that's the spot to add that? | 22:01 |
| comstud | vishy: Yeah, I wanted to ask you about that, too | 22:01 |
| vishy | comstud: the idea is that we have some arbitrary data passed in by the user that makes it to the scheduler | 22:01 |
| vishy | and this data can be interpreted by the filter or weighter | 22:01 |
| comstud | i assume that mostly that's admin-only | 22:02 |
| _0x44 | Not all of it should be admin-only | 22:02 |
| vishy | nova boot --scheduler_hints = "{'different_host': 'instance:<uuid>'}' | 22:02 |
| vishy | i.e. make sure it is on a different host than this other instance | 22:03 |
| *** lloydde has joined #openstack-dev | 22:03 | |
| vishy | if the filter is enabled it would make sure you got a different host | 22:03 |
| comstud | Okay, right | 22:03 |
| vishy | in this case it would be a user facing feature | 22:03 |
| vishy | the affinity one was just kind of experimental | 22:03 |
| vishy | i don't know if same_host is useful | 22:03 |
| comstud | well, we have affinity on our list | 22:03 |
| vishy | but different_host is for sure | 22:04 |
| openstackgerrit | Verification of a change to openstack/glance failed: Bug #909574: Glance does not sanity-check given image size on upload https://review.openstack.org/2727 | 22:04 |
| uvirtbot | Launchpad bug 909574 in glance "Glance does not sanity-check given image size on upload" [Medium,In progress] https://launchpad.net/bugs/909574 | 22:04 |
| comstud | but affinity essentially maps toa zone for us | 22:04 |
| comstud | not a host | 22:04 |
| comstud | as far as host, i get it | 22:04 |
| vishy | comstud: yeah i would think host_aggregates might get you there | 22:04 |
| vishy | comstud: if availability zones don't work for some reason | 22:04 |
| Vek | can I get mtaylor or jeblair to look at https://review.openstack.org/#change,3190 please? devstack keeps failing to start nova-api, but I can't replicate and need more information. | 22:05 |
| vishy | comstud: anyway, bcwaldon was going to add the extension to allow scheduler hints | 22:05 |
| comstud | you coming to SAT 2/1 ? | 22:05 |
| vishy | comstud: it appears so | 22:05 |
| comstud | this might be a topic for discussion | 22:05 |
| vishy | comstud: I need the different_host scheduler in prior to that, hence the rewrite work | 22:05 |
| vishy | comstud: but for the larger affinity question, for sure | 22:06 |
| comstud | nod | 22:06 |
| comstud | yea | 22:06 |
| comstud | well, i was going to take care of part of waldon's work | 22:06 |
| comstud | he can do the extension | 22:06 |
| comstud | I can get the framework in place | 22:06 |
| *** lts has quit IRC | 22:06 | |
| comstud | i have part of it done | 22:07 |
| *** zigo has quit IRC | 22:08 | |
| comstud | in fact, I think i just have tests to fix | 22:09 |
| *** david-kranz has quit IRC | 22:09 | |
| bcwaldon | comstud: feel free to do it | 22:10 |
| comstud | what I'm doing is passing filter_properties dict to scheduler from compute | 22:10 |
| comstud | you can add what you need there | 22:10 |
| comstud | when i'm done | 22:10 |
| bcwaldon | ok | 22:11 |
| bcwaldon | just give me your branch so I can see what you're doing | 22:11 |
| comstud | yeah | 22:11 |
| comstud | i'll push something up | 22:11 |
| *** bencherian has joined #openstack-dev | 22:12 | |
| bcwaldon | comstud: weren't you supposed to be off today? | 22:13 |
| comstud | no | 22:13 |
| bcwaldon | yes? | 22:14 |
| comstud | no | 22:15 |
| comstud | originally i was going to take a few days off | 22:15 |
| comstud | but decided I didn't need to right now | 22:15 |
| comstud | fg | 22:15 |
| comstud | oops | 22:15 |
| *** byeager has joined #openstack-dev | 22:17 | |
| bcwaldon | bg | 22:18 |
| *** adjohn has quit IRC | 22:19 | |
| openstackgerrit | Florian Hines proposed a change to openstack/swift: swift-recon cli cleanup https://review.openstack.org/2928 | 22:19 |
| *** adjohn has joined #openstack-dev | 22:19 | |
| comstud | bcwaldon: https://github.com/comstud/nova/commit/639f912d1e79d70cf5aae56e9a38030fd3d99a99 | 22:20 |
| *** mdomsch_ has quit IRC | 22:21 | |
| bcwaldon | comstud: ok, and you want a dictionary from the user to update the defaults you have set? | 22:21 |
| comstud | i would add new entries into filter_properties | 22:22 |
| comstud | like: | 22:22 |
| comstud | 'desired_host': <whatever> | 22:22 |
| bcwaldon | ok, but you've got some already defined, like local_zone_only | 22:22 |
| bcwaldon | are those mutable by the user? | 22:22 |
| comstud | that should not be | 22:22 |
| comstud | that'll be going away | 22:22 |
| comstud | ignore the zones stuff | 22:23 |
| *** bsza has quit IRC | 22:23 | |
| comstud | ignore_hosts right now is only set by the compute manager for a resize based on a flag | 22:23 |
| comstud | it's empty for an initial build, so you could use it | 22:23 |
| bcwaldon | ok, I'll see what I can do | 22:23 |
| comstud | how are you going to handle rebuilds and resizing which do re-scheduling? | 22:23 |
| comstud | I guess you'll have to support that in extension too | 22:24 |
| bcwaldon | yeah, same interface I use for create | 22:24 |
| bcwaldon | you buy that? | 22:24 |
| comstud | yeah | 22:24 |
| bcwaldon | is the branch 'filter_changes'? | 22:24 |
| comstud | feel free to use ignore_hosts | 22:24 |
| comstud | yeah | 22:25 |
| comstud | i think this is ready for review | 22:25 |
| bcwaldon | how would I 'use' it? | 22:25 |
| Vek | can I get mtaylor or jeblair to look at https://review.openstack.org/#change,3190 please? devstack keeps failing to start nova-api, but I can't replicate and need more information. | 22:25 |
| *** rnirmal has quit IRC | 22:25 | |
| bcwaldon | comstud: not sure what you mean | 22:25 |
| comstud | if you find that you want to allow a user to ignore certian hosts, feel free to take advantage of the current 'ignore_hosts' entry in the filter properties | 22:25 |
| bcwaldon | ok, I'm not going to do any validation | 22:26 |
| bcwaldon | the keys need to be open so any scheduler implementation can get what it needs | 22:26 |
| *** cp16net has quit IRC | 22:26 | |
| comstud | if you want to add a 'build only on this host' thing, you'll need to add a new key to the filter_properties dict | 22:26 |
| comstud | yes | 22:26 |
| bcwaldon | so the keys will be defined by the scheduler | 22:26 |
| bcwaldon | or whatever filters are supported | 22:27 |
| comstud | yeah, filters typically | 22:27 |
| bcwaldon | but I won't worry about that at my level | 22:27 |
| comstud | the only hard coded one is 'ignore_hosts' (and 'local_zone_only' for zone stuff) | 22:27 |
| comstud | i could move that into a filter | 22:27 |
| bcwaldon | sure, they'r hard-coded in the compute layer, but I'm still going to treat them as arbitrary keys in the api | 22:27 |
| comstud | nod | 22:28 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Implement Secure Token Auth https://review.openstack.org/2889 | 22:29 |
| *** bencherian has quit IRC | 22:30 | |
| comstud | fixed tests.. running now to make sure | 22:36 |
| *** bencherian has joined #openstack-dev | 22:38 | |
| openstackgerrit | Verification of a change to openstack/keystone failed: Addresses bug 918608 https://review.openstack.org/3202 | 22:40 |
| openstackgerrit | sleepsonthefloor proposed a change to openstack/horizon: Port quotas to novaclient https://review.openstack.org/3240 | 22:41 |
| openstackgerrit | Chris Behrens proposed a change to openstack/nova: pass filter_properties into scheduling requests for run_instance https://review.openstack.org/3241 | 22:41 |
| comstud | bcwaldon: ^^ | 22:42 |
| bcwaldon | comstud: kk, working on the extension meow | 22:42 |
| bcwaldon | comstud: vg break | 22:42 |
| comstud | vg break? | 22:42 |
| bcwaldon | video game | 22:42 |
| bcwaldon | or is it | 22:42 |
| comstud | i see | 22:43 |
| openstackgerrit | sleepsonthefloor proposed a change to openstack/horizon: Port servers calls to novaclient https://review.openstack.org/3242 | 22:44 |
| Vek | can I get mtaylor or jeblair to look at https://review.openstack.org/#change,3190 please? devstack keeps failing to start nova-api, but I can't replicate and need more information. | 22:46 |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Restore Console Info Logging - bp keystone-logging https://review.openstack.org/3243 | 22:47 |
| openstackgerrit | Aaron Lee proposed a change to openstack/melange: Adding pip install instructions to readme. https://review.openstack.org/3244 | 22:49 |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Implement Secure Token Auth https://review.openstack.org/2889 | 22:49 |
| openstackgerrit | Joe Gordon proposed a change to openstack/nova: Pull out ram_filter into a separate filter https://review.openstack.org/3211 | 22:53 |
| *** adjohn has quit IRC | 22:56 | |
| *** adjohn has joined #openstack-dev | 22:56 | |
| openstackgerrit | Verification of a change to openstack/keystone failed: Implement Secure Token Auth https://review.openstack.org/2889 | 23:00 |
| openstackgerrit | Ziad Sawalha proposed a change to openstack/keystone: Addresses bug 918608 https://review.openstack.org/3202 | 23:00 |
| *** bencherian has quit IRC | 23:01 | |
| *** troytoman-away is now known as troytoman | 23:03 | |
| openstackgerrit | dlapsley proposed a change to openstack/nova: bug 917397 https://review.openstack.org/3245 | 23:08 |
| uvirtbot | Launchpad bug 917397 in nova/essex "Quantum Manager configuration sanity checks" [Undecided,New] https://launchpad.net/bugs/917397 | 23:08 |
| *** bencherian has joined #openstack-dev | 23:13 | |
| Vek | can I get mtaylor or jeblair to look at https://review.openstack.org/#change,3190 please? devstack keeps failing to start nova-api, but I can't replicate and need more information. | 23:14 |
| *** LinuxJedi has quit IRC | 23:18 | |
| openstackgerrit | Jake Dahn proposed a change to openstack/python-novaclient: Implementing client for new x509 support in nova. https://review.openstack.org/3246 | 23:20 |
| openstackgerrit | Dave Lapsley proposed a change to openstack/nova: bug 917397 https://review.openstack.org/3245 | 23:22 |
| uvirtbot | Launchpad bug 917397 in nova/essex "Quantum Manager configuration sanity checks" [Undecided,New] https://launchpad.net/bugs/917397 | 23:22 |
| openstackgerrit | Rick Harris proposed a change to openstack/nova: Fix xenapi rescue without swap. https://review.openstack.org/3247 | 23:23 |
| *** crobinso has quit IRC | 23:25 | |
| comstud | vishy: ping | 23:26 |
| comstud | vishy: I'm going to push up test fixes for your branch, and rebase it off another review... i think it can go in and you can move some things around. nice to get it in so request_spec, context, etc is there. | 23:29 |
| comstud | let's see if this works | 23:29 |
| openstackgerrit | Chris Behrens proposed a change to openstack/nova: Adds Availability Zone Filter https://review.openstack.org/3214 | 23:30 |
| vishy | comstud: i'm just splitting it into multiple patches | 23:31 |
| comstud | hm, it didn't grab my changes anyway | 23:31 |
| comstud | ok | 23:31 |
| vishy | trying to remember how to get git to stage part of a file | 23:31 |
| *** jakedahn has quit IRC | 23:32 | |
| comstud | no idea, i'm not a git person | 23:32 |
| *** jakedahn has joined #openstack-dev | 23:32 | |
| comstud | vishy: If it saves you time, you can steal this: https://github.com/comstud/nova/commit/5aa8bd4ac7f25d42912f7cf897fa6df3e8a14995#diff-4 | 23:35 |
| comstud | the dist sched test fixes | 23:35 |
| comstud | if it doesn't save you time, no big deal | 23:35 |
| comstud | :) | 23:35 |
| Vek | vishy: for staging parts of files, I usually use git gui. There is a command-line way of doing it, but I'm not sure what; check "git help add" maybe? | 23:36 |
| vishy | git add --patch | 23:37 |
| vishy | or git add -i | 23:37 |
| vishy | pretty awesome | 23:37 |
| Vek | *nod* I use that feature in git gui all the time. | 23:38 |
| openstackgerrit | Rick Harris proposed a change to openstack/nova: ERROR out instance if unrescue fails. https://review.openstack.org/3248 | 23:40 |
| Vek | vishy: btw, one more patch for the extension-refactor blueprint, but I can't get it past devstack and I can't figure out why. | 23:41 |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Add option to force hosts to scheduler https://review.openstack.org/3249 | 23:41 |
| comstud | vishy: Hm... your patch would be much nicer after mine goes in | 23:46 |
| comstud | vishy: see: https://review.openstack.org/#change,3241 | 23:46 |
| comstud | vishy: You don't need to add a new argument to run_instance(). You can just add force_hosts to filter_properties | 23:47 |
| vishy | hmm | 23:47 |
| vishy | that would be nicer | 23:47 |
| comstud | take a look, approve, and rebase! | 23:47 |
| comstud | ;) | 23:47 |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Add context and request spec to filter_properties https://review.openstack.org/3250 | 23:48 |
| comstud | that will need a rebase if we can get an Approved on mine as well | 23:49 |
| comstud | and then you can ditch populate_filter_properties completely | 23:49 |
| vishy | ok those are my three patches | 23:50 |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Adds availability zone filter https://review.openstack.org/3251 | 23:50 |
| vishy | all separate and tests passing isolated on each one | 23:50 |
| vishy | well except for pep8 of course | 23:50 |
| vishy | :( | 23:50 |
| comstud | hah | 23:50 |
| openstackgerrit | Vish Ishaya proposed a change to openstack/nova: Adds availability zone filter https://review.openstack.org/3251 | 23:51 |
| openstackgerrit | andrewbogott proposed a change to openstack/nova: Explicitly encode string to utf8 before passing to ldap. https://review.openstack.org/3252 | 23:51 |
| vishy | cool | 23:52 |
| vishy | i can add it in to filter properties instead. | 23:52 |
| comstud | yeah, it's a bit more clean if compute passes it in | 23:53 |
| comstud | then waldon can do his extension work | 23:53 |
| comstud | and your force_hosts work is easy | 23:54 |
| comstud | sorry, kinda stomped on each other here, but easy to resolve | 23:54 |
| *** troytoman is now known as troytoman-away | 23:54 | |
| vishy | ok just approved yours | 23:54 |
| comstud | thnx | 23:54 |
| vishy | will rebase once it merges | 23:54 |
| vishy | i think az filter doesn't need a rebase | 23:54 |
| comstud | looking | 23:55 |
| vishy | comstud: hmm but it probably needs one of my other branches to actually work | 23:56 |
| *** Gordonz has quit IRC | 23:56 | |
| comstud | yeah it will | 23:56 |
| comstud | i approved anyway.. i guess i should not have since you added it to default filter list | 23:57 |
| comstud | er | 23:57 |
| comstud | +2'd | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!