| *** stiekes_ has quit IRC | 00:42 | |
| *** stiekes_ has joined #openstack-dev | 00:42 | |
| *** medberry is now known as med_out | 01:02 | |
| openstackjenkins | Project nova build #1,214: SUCCESS in 3 min 14 sec: http://jenkins.openstack.org/job/nova/1214/ | 01:59 |
|---|---|---|
| openstackjenkins | Tarmac: Adds ability to disable snapshots in the Openstack API. | 01:59 |
| *** tyoc213 has joined #openstack-dev | 02:25 | |
| *** stiekes has joined #openstack-dev | 02:58 | |
| *** stiekes_ has quit IRC | 02:58 | |
| *** stiekes_ has joined #openstack-dev | 03:19 | |
| *** stiekes has quit IRC | 03:21 | |
| *** stiekes has joined #openstack-dev | 03:23 | |
| *** stiekes_ has quit IRC | 03:24 | |
| *** nci has quit IRC | 03:32 | |
| *** nci has joined #openstack-dev | 03:33 | |
| *** tyoc213 has quit IRC | 04:03 | |
| *** martine has quit IRC | 04:17 | |
| *** westmaas has quit IRC | 04:24 | |
| *** westmaas has joined #openstack-dev | 04:24 | |
| *** winston-d has quit IRC | 05:55 | |
| *** jakedahn has joined #openstack-dev | 05:59 | |
| *** nickon has joined #openstack-dev | 06:35 | |
| *** zigo has joined #openstack-dev | 07:28 | |
| *** zigo is now known as zigo-_- | 07:34 | |
| *** tudamp has joined #openstack-dev | 07:43 | |
| *** zul has joined #openstack-dev | 07:54 | |
| *** TimR-L has joined #openstack-dev | 08:13 | |
| *** alekibango has joined #openstack-dev | 08:44 | |
| *** darraghb has joined #openstack-dev | 09:23 | |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/keystone: Add proper simple_bind_s to fakeldap, removed all imports from ldap. https://review.openstack.org/159 | 09:27 |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/keystone: Add LDAP schema. https://review.openstack.org/163 | 09:27 |
| *** phil121 has joined #openstack-dev | 10:43 | |
| *** bourke has quit IRC | 11:00 | |
| *** markvoelker has joined #openstack-dev | 11:04 | |
| *** martine has joined #openstack-dev | 11:50 | |
| *** martine has quit IRC | 11:59 | |
| *** mfer has joined #openstack-dev | 12:03 | |
| *** mnour has joined #openstack-dev | 12:04 | |
| *** lorin1 has joined #openstack-dev | 12:14 | |
| *** lorin1 has quit IRC | 12:23 | |
| *** lorin1 has joined #openstack-dev | 12:29 | |
| *** dprince has joined #openstack-dev | 12:43 | |
| *** lts has joined #openstack-dev | 12:54 | |
| lts | Hi, Gaitan, Ndk: I'm back. Bad power hits here yesterday. I have one unrecoverable lv on my z400. Moving mail over to my notebook so I can stay in touch while I work on the z400. | 12:55 |
| lts | Fortunately, the replacement drive arrived just before vacation, so I can install that and recover what I can from the drive and restore the rest from backups or recreate. I've lost a few VM images, but nothing important. | 12:56 |
| *** zigo-_- has quit IRC | 13:04 | |
| *** zigo has joined #openstack-dev | 13:04 | |
| *** bcwaldon has joined #openstack-dev | 13:20 | |
| *** blamar has quit IRC | 13:28 | |
| *** blamar_ has joined #openstack-dev | 13:29 | |
| *** bcwaldon has quit IRC | 13:30 | |
| *** bcwaldon has joined #openstack-dev | 13:30 | |
| *** bsza has joined #openstack-dev | 13:34 | |
| *** kbringard has joined #openstack-dev | 13:34 | |
| *** bsza has quit IRC | 13:37 | |
| *** bsza has joined #openstack-dev | 13:37 | |
| *** amccabe has joined #openstack-dev | 13:37 | |
| *** vladimir3p has joined #openstack-dev | 13:41 | |
| *** jshepher has joined #openstack-dev | 13:48 | |
| *** zigo is now known as zigo-_- | 13:50 | |
| *** rnirmal has joined #openstack-dev | 14:02 | |
| *** mattray has joined #openstack-dev | 14:14 | |
| *** dolphm has joined #openstack-dev | 14:23 | |
| *** zigo-_- has quit IRC | 14:25 | |
| *** Binbin has joined #openstack-dev | 14:26 | |
| *** chomping has joined #openstack-dev | 14:30 | |
| *** zigo has joined #openstack-dev | 14:30 | |
| *** jkoelker has joined #openstack-dev | 14:33 | |
| *** dolphm has quit IRC | 14:33 | |
| *** nickon has quit IRC | 14:35 | |
| tudamp | sandywalsh: hi, you have a minute? | 14:36 |
| sandywalsh | tudamp, yup | 14:37 |
| tudamp | ok, you think this scenario: compute configured with raw images, so the instances take a few minutes to be up. | 14:38 |
| tudamp | 1) launch an instance; | 14:38 |
| tudamp | 2) scheduler consumes resources on host_list | 14:38 |
| tudamp | 3) before the instance is up compute update scheduler | 14:38 |
| *** dolphm has joined #openstack-dev | 14:38 | |
| tudamp | 4) resources consumed are overwritten | 14:38 |
| tudamp | if you launch an instance now I will use the resources occupied by the instance that is going to get up | 14:40 |
| *** nmistry has joined #openstack-dev | 14:40 | |
| sandywalsh | tudamp, correct, there will be a lag between reported resources and provisioning in-progress | 14:40 |
| *** zigo has quit IRC | 14:41 | |
| tudamp | could be avoided by updating the scheduler every time an instance is migrated, turned off, etc. | 14:41 |
| sandywalsh | tudamp, it's really an issue when provisioning 1-off instances (server X, then server Y, then server Z, etc) | 14:41 |
| sandywalsh | tudamp, perhaps less of an issue if you say "provision N servers" ... since the CostScheduler "consumes" resources when allocating hosts | 14:42 |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/keystone: Add wrapper for real LDAP connection with logging and type converting. https://review.openstack.org/162 | 14:42 |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/keystone: Fix console and debug logging. https://review.openstack.org/161 | 14:42 |
| openstackgerrit | Yuriy Taraday proposed a change to openstack/keystone: Moved password check logic to backend https://review.openstack.org/160 | 14:42 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Fix console and debug logging. https://review.openstack.org/161 | 14:42 |
| *** cp16net has joined #openstack-dev | 14:42 | |
| sandywalsh | tudamp, yes, we could be smarter about when we send our updates | 14:42 |
| sandywalsh | tudamp, but ... the schedulers may go up and down ... the periodic updates are to inform the new ones of the state of the world | 14:43 |
| *** med_out is now known as medberry | 14:44 | |
| tudamp | sandywalsh: mm.. therefore, also in this case, the scheduler is not completely reliable | 14:47 |
| openstackjenkins | Project swift build #304: SUCCESS in 34 sec: http://jenkins.openstack.org/job/swift/304/ | 14:47 |
| openstackjenkins | Tarmac: PEP8 Fix | 14:47 |
| *** nmistry has quit IRC | 14:59 | |
| sandywalsh | tudamp, well, it's as reliable as the data it has. | 15:00 |
| sandywalsh | tudamp, we have to assume N scheduler and M compute nodes | 15:01 |
| tudamp | sandywalsh, I have not seen LeastCostScheduler :p | 15:06 |
| tudamp | sandywalsh, I thought that the scheduler update was to keep synchronized nodes / zones / scheduler | 15:07 |
| tudamp | sandywalsh, and not for inform the new ones of the state of the world | 15:08 |
| sandywalsh | tudamp, there are two updates that occur. 1. the ZoneManager in each Scheduler pings the child zones for status updates and 2. The services within a Zone fanout update the schedulers with the nodes state | 15:08 |
| sandywalsh | tudamp, But, I'm still not really clear what your question is? | 15:09 |
| *** nickon has joined #openstack-dev | 15:10 | |
| tudamp | sandywalsh: I just want that an instance go running only on a host that can accommodate :) | 15:12 |
| tudamp | sandywalsh: easy to say, less easy to do :) | 15:13 |
| sandywalsh | tudamp, the LeastCostScheduler is what you want. We're cleaning it up now for our alpha release. | 15:15 |
| sandywalsh | tudamp, there was some confusion on filter_hosts() and weigh_hosts(), which is also getting straightened up. | 15:15 |
| tudamp | sandywalsh: ok, then i look for updates. | 15:18 |
| tudamp | sandywalsh: do not read criticism in my statements, I'm just trying to figure out how it works and if I can be useful for something. | 15:18 |
| sandywalsh | tudamp, absolutely ... we just have the framework in place, now we need to turn the knobs to get it working properly. Any testing/changes you care to make would be most welcome. | 15:19 |
| *** martine has joined #openstack-dev | 15:23 | |
| *** Binbin has quit IRC | 15:23 | |
| tudamp | sandywalsh: great, I thought to use your framework | 15:24 |
| tudamp | sandywalsh: to create a time-scheduler, with advanced resource reservations. But this will not happen before October | 15:25 |
| tudamp | sandywalsh: Before LeastCost must work, is what we need now. We will be happy to test it. | 15:25 |
| sandywalsh | tudamp, sounds great ... let me know how I can help. I don't think we're far off from getting some basic (real) scheduling in place. | 15:26 |
| tudamp | sandywalsh: today I tried the zoneawarescheduler: without the bad data coming from the nodes seems to do its job ;) | 15:31 |
| *** rnorwood has joined #openstack-dev | 15:32 | |
| sandywalsh | zone aware scheduler by itself won't do much ... it's a base class for real schedulers. HostFilterScheduler is perhaps the most basic scheduler to derive from it | 15:33 |
| *** ameade has joined #openstack-dev | 15:38 | |
| *** mnour has quit IRC | 15:48 | |
| *** dprince has quit IRC | 15:52 | |
| *** mdomsch has quit IRC | 15:54 | |
| openstackgerrit | Dolph Mathews proposed a change to openstack/keystone: Add proper simple_bind_s to fakeldap, removed all imports from ldap. https://review.openstack.org/159 | 15:54 |
| jk0 | unit tests in trunk failing for anyone? | 15:56 |
| jk0 | ah, I see why | 15:57 |
| dolphm | which project? | 15:59 |
| jk0 | nova | 16:01 |
| jk0 | but I figured it out | 16:01 |
| *** yogirackspace has joined #openstack-dev | 16:03 | |
| sandywalsh | ttx soren mtaylor ... need novaclient update so this can go through https://code.launchpad.net/~sandy-walsh/nova/v1_1_novaclient/+merge/70474 | 16:09 |
| *** Transformer has joined #openstack-dev | 16:13 | |
| *** tudamp has left #openstack-dev | 16:14 | |
| openstackgerrit | Dolph Mathews proposed a change to openstack/keystone: Add proper simple_bind_s to fakeldap, removed all imports from ldap. https://review.openstack.org/159 | 16:16 |
| *** dragondm has joined #openstack-dev | 16:16 | |
| *** Transformer has left #openstack-dev | 16:17 | |
| *** erik-s has quit IRC | 16:18 | |
| *** erik-s has joined #openstack-dev | 16:18 | |
| openstackgerrit | A change was merged to openstack/keystone: Add proper simple_bind_s to fakeldap, removed all imports from ldap. https://review.openstack.org/159 | 16:18 |
| mtaylor | sandywalsh: ok. just got the slave updated | 16:25 |
| sandywalsh | thanks mtaylor, we'll reapprove | 16:26 |
| *** zul has quit IRC | 16:32 | |
| *** jakedahn has quit IRC | 16:34 | |
| vladimir3p | jaypipes: ping | 16:35 |
| jaypipes | vladimir3p: pong | 16:35 |
| vladimir3p | jaypipes: trying to merge latest nova code and getting a pretty strange error that seems to be glance related | 16:36 |
| vladimir3p | jaypipes: http://paste.openstack.org/show/2119/ | 16:36 |
| sandywalsh | devcamcar, jk0 seems I need a reapprove on https://code.launchpad.net/~sandy-walsh/nova/v1_1_novaclient/+merge/70764 | 16:36 |
| jaypipes | vladimir3p: looking... | 16:36 |
| jaypipes | vladimir3p: have you upgrade glance? | 16:37 |
| jk0 | sandywalsh: looks like you might need to submit a new MP, it's not letting me reapprove | 16:37 |
| vladimir3p | jaypipes: it should be on diablo-3 release | 16:37 |
| jaypipes | vladimir3p: what does glance --version say? | 16:37 |
| vladimir3p | jaypipes: glance 2011.3-dev | 16:38 |
| jaypipes | vladimir3p: hmm... | 16:38 |
| jaypipes | Vek: around? | 16:38 |
| jaypipes | vladimir3p: one sec... | 16:38 |
| sandywalsh | jk0, just re-proposed | 16:39 |
| jk0 | sandywalsh: still wont let me. looks like you need to delete this one and make a new MP | 16:39 |
| sandywalsh | phooey | 16:39 |
| *** zaitcev has joined #openstack-dev | 16:40 | |
| sandywalsh | jk0, https://code.launchpad.net/~sandy-walsh/nova/v1_1_novaclient/+merge/70765 | 16:40 |
| *** jdurgin has joined #openstack-dev | 16:40 | |
| dabo | Still looking for some nova core reviews for https://code.launchpad.net/~ed-leafe/nova/powerstate/+merge/70230 | 16:41 |
| s1rp | vladimir3p: i ran into that error too, i think python setup.py install on glance fixed it for me | 16:41 |
| vladimir3p | s1rp: thanks will try it right now | 16:42 |
| s1rp | vladimir3p: and make sure you bounce nova after you do that | 16:42 |
| *** medberry is now known as med_out | 16:46 | |
| *** dprince has joined #openstack-dev | 16:59 | |
| openstackgerrit | Jay Pipes proposed a change to openstack/glance: Add @skip_if_disabled decorator to test.utils and integrate it into the base functional API test case. The S3 functional test case now uses test_api.TestApi as its base class and the setUp() method sets the disabled and disabled_message attributes that th https://review.openstack.org/177 | 16:59 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Fix console and debug logging. https://review.openstack.org/161 | 17:01 |
| dprince | mtaylor: you around? | 17:10 |
| mtaylor | dprince: morning! | 17:10 |
| dprince | mtaylor: Hey. So is there a good way I get get a list of branches up for review from gerrit? | 17:11 |
| dprince | mtaylor: Preferably on the command line. | 17:11 |
| mtaylor | dprince: perhaps https://review.openstack.org/Documentation/cmd-query.html | 17:12 |
| dprince | mtaylor: In a human readable branch form? I see gerrit is using Refspec's in the webUI. | 17:12 |
| mtaylor | hrm... jeblair ^^^ any thoughts? | 17:13 |
| dprince | mtaylor: Ideally I'd like to be able to fetch someones code w/ the same they used locally for the branch. | 17:13 |
| mtaylor | dprince: I don't know off the top of my head - I'm _sure_ there is a way | 17:13 |
| mtaylor | dprince: ah, well, now that I know can be done | 17:13 |
| mtaylor | dprince: (there's a fetch thing in the web ui - just a matter of finding the right thing in the command line tool) ... lemme poke | 17:14 |
| dprince | mtaylor: Okay. Well I'm looking for that. I'd like to make it easy for humans to test branches up for review on Gerrit. | 17:14 |
| dprince | mtaylor: Using the refspecs will work but man oh man that is going to be a lot of detective work to keep up with that. | 17:15 |
| mtaylor | dprince: you've seen the fetch/pull lines on the reviews themselves, yeah? | 17:15 |
| dprince | mtaylor: Yes. They all just use the refspecs. For example 'refs/changes/70/170/1' | 17:16 |
| dprince | I'd like to pull the branch name with something a bit more humanly understandable. Like 'fix_osapi_server_create' or something like that. | 17:17 |
| mtaylor | dprince: oh, I do not believe that exists | 17:17 |
| mtaylor | dprince: I'd want to verify with jeblair to make sure - but gerrit does magic with refspecs in terms of internal storage | 17:18 |
| jeblair | ssh review gerrit query "status: open" --current-patch-set | 17:18 |
| dprince | jeblair: Thanks. Let me check that out. | 17:18 |
| jeblair | gets a list of the current open patches, including their refspcs | 17:19 |
| *** bcwaldon has quit IRC | 17:19 | |
| jeblair | ssh review gerrit query "status: open" --current-patch-set --format JSON | 17:19 |
| jeblair | dprince: ^ that could be the basis of a script that does what you want | 17:19 |
| jeblair | that gets you patches, their descriptions, and refspecs | 17:19 |
| jaypipes | JordanRinke: around? want to know what your opinion is on the priority of https://bugs.launchpad.net/glance/+bug/819936? | 17:21 |
| uvirtbot | Launchpad bug 819936 in glance "New image cache breaks Glance on Windows" [Low,Confirmed] | 17:21 |
| jaypipes | JordanRinke: seems to me that having the image cache be truly a disable-able middleware/extension will fix things for Windows, right? | 17:22 |
| sandywalsh | any keystone people in the house? | 17:25 |
| dprince | jeblair: Okay. That command will give me 'subject' descriptions which I can then associate w/ the refspecs. That is better but I'd love it if we could use humanly readable branch names instead. | 17:25 |
| dolphm | sandywalsh: <-- | 17:25 |
| sandywalsh | dolphm, hey! trying to test novaclient integration with keystone | 17:25 |
| sandywalsh | dolphm, getting http://paste.openstack.org/show/2120/ | 17:26 |
| sandywalsh | dolphm, the service is running | 17:26 |
| *** jakedahn has joined #openstack-dev | 17:26 | |
| jeblair | dprince: i think subject is the closest thing that gerrit provides. it picks it up from the first line of the commit. | 17:26 |
| sandywalsh | dolphm, btw, seems --host / --bind-host don't work | 17:26 |
| sandywalsh | dolphm, (it's listening on 0.0.0.0 by default) | 17:27 |
| dolphm | hmm | 17:27 |
| dolphm | sandywalsh: on the --host issue, feel free to open an issue on github.com/rackspace/keystone | 17:28 |
| sandywalsh | dolphm, yup, was just going to | 17:28 |
| *** alekibango has quit IRC | 17:28 | |
| dprince | jeblair/mtaylor: Another question. The Anonymous HTTP commands don't work for me. | 17:29 |
| dolphm | sandywalsh: cool, as for the x-server-management-url, i've seen that once before, but was unable to reproduce | 17:29 |
| dolphm | sandywalsh: it popped up in our test suite, and then disappeared | 17:29 |
| *** darraghb has quit IRC | 17:29 | |
| sandywalsh | hmm | 17:29 |
| dolphm | sandywalsh: but it happens every time for you? | 17:30 |
| dprince | jeblair/mtaylor: So if I click on the 'Download' --> 'Anonymous HTTP' link and try to run that command in my cloned project it fails. | 17:30 |
| sandywalsh | yeah, seems to | 17:30 |
| sandywalsh | dolphm, I'll put some debugging in there to see what's happening | 17:30 |
| dprince | jeblair/mtaylor: The SSH commands do appear to work (although they'd require me SSH key to be everywhere I'd need to run them). | 17:30 |
| jeblair | dprince: what error do you get? | 17:31 |
| *** bcwaldon has joined #openstack-dev | 17:31 | |
| dprince | fatal: HTTP request failed | 17:31 |
| mtaylor | dprince, jeblair: ssl cert | 17:32 |
| mtaylor | "error: server certificate verification failed. CAfile: /etc/ssl/certs/ca-certificates.crt CRLfile: none while accessing https://review.openstack.org/p/openstack/glance/info/refs" | 17:32 |
| mtaylor | jeblair: ^^ | 17:32 |
| jeblair | mtaylor: thanks | 17:33 |
| openstackgerrit | Vish Ishaya proposed a change to openstack/keystone: Adds support for authenticating via ec2 signatures https://review.openstack.org/127 | 17:34 |
| dolphm | sandywalsh: i'll forward this to the yogi, who i believe wrote the related code in keystone | 17:34 |
| sandywalsh | dolphm, cool, thanks. I'll let you know what I find. | 17:35 |
| dolphm | sandywalsh: or if you want to go ahead and open an issue on this too, i'll make sure it's assigned to him | 17:36 |
| sandywalsh | dolphm, will do. (odd, it's getting a 305 so something is answering) | 17:36 |
| *** bcwaldon has quit IRC | 17:36 | |
| dolphm | i *think* the rest of the response is "correct" but it shouldn't include a stack trace - the 305 is indicating that you're trying to call nova without first authenticating via the indicated url (keystone) | 17:37 |
| *** bcwaldon has joined #openstack-dev | 17:37 | |
| dolphm | sandywalsh: the request contains an "x-auth-key" headers, which i'm not familiar with... keystone uses the "x-auth-token" header... but that may be what the specific test is testing for | 17:39 |
| dolphm | sandywalsh: ... or is that not a test? lol | 17:40 |
| sandywalsh | dolphm, we have two merge-props in novaclient for keystone integration ... let me see if the other assembles the same request | 17:40 |
| sandywalsh | both are specifying x-auth-key ... lemme check what keystone is looking for (in my version) | 17:41 |
| *** Tushar has joined #openstack-dev | 17:41 | |
| Tushar | I am trying to setup zone_aware_scheduler to test launching of server on specific host. But I didn't find the format of blob that I should be passing to the Create Server API. Does anyone knows the format of blob? | 17:43 |
| sandywalsh | Tushar, you don't need to parse the blob. It gets created by the /zones/select command | 17:50 |
| *** jakedahn has quit IRC | 17:50 | |
| Tushar | sandywalsh: so before calling /zones/boot I should make a call to /zones/select. Is it correct? | 17:52 |
| sandywalsh | Tushar, you should nova.scheduler.host_filter:HostFilterScheduler as your base zone-aware scheduler (or LeastCostScheduler) | 17:52 |
| sandywalsh | Tushar, no need ... the framework will do that for you. You can just call zone-boot | 17:52 |
| sandywalsh | Tushar, 'nova zone-boot servername --image X --flavor Y' | 17:53 |
| sandywalsh | Tushar, it'll call /zone/select and deal with all the blob stuff under the hood | 17:53 |
| Tushar | sandywalsh: I am specifically interested to launch the vm instance on specific host. Is that possible using zone_ware_scheduler? | 17:53 |
| kbringard | you'd probably have to create a single compute node zone | 17:53 |
| kbringard | then send it to that zone | 17:54 |
| sandywalsh | Tushar, not directly, the purpose of the scheduler is to pick the best host based on the request parameters. If your request was so specific that there was only one host for the job, that might do it. | 17:54 |
| *** heckj has joined #openstack-dev | 17:54 | |
| sandywalsh | Tushar, but right now we only pick hosts based on InstanceType (memory and disk) | 17:55 |
| sandywalsh | Tushar, I know there was work going on to allow extra parameters on instance type, but not sure the status of that | 17:55 |
| Tushar | sandywalsh: I am referring to latest trunk code and trying to understand method "_provision_resource_locally" | 17:57 |
| jk0 | jenkins backed up? | 17:57 |
| Tushar | kbringard: Thanks. I got it. | 17:57 |
| sandywalsh | Tushar, that method is called when it has been decided to provision the instance in the current zone (locally). It creates a db entry for the instance and forwards the request to the appropriate compute node. | 17:58 |
| kbringard | in the future you could probably set a zone_capability attribute for each compute node | 17:58 |
| sandywalsh | kbringard, agreed | 17:59 |
| kbringard | once the scheduler can key off of any of those attributes | 17:59 |
| openstackgerrit | Brian Waldon proposed a change to openstack/glance: breaking up MAX_ITEM_LIMIT and making the new values configurable https://review.openstack.org/170 | 18:00 |
| Tushar | sandywalsh: I am still not very clear about how to pass blob to the create server API. If I don't pass blob then I think it will raise InvalidBlob error. | 18:03 |
| jeblair | dprince: still here? | 18:04 |
| jeblair | dprince: try the https git fetch command again | 18:04 |
| *** stiekes has quit IRC | 18:05 | |
| jeblair | dprince,mtaylor: the ssl certs we got for jenkins.o.o and review.o.o had _two_ intermediate certs, but apache was only serving one of them. i fixed that, and it works for me now. | 18:05 |
| *** stiekes has joined #openstack-dev | 18:05 | |
| sandywalsh | Tushar, you don't need to ... the scheduler will append it when it forwards a boot request to a child zone | 18:07 |
| sandywalsh | Tushar, if the child zone sees a blob it will attempt to use it, otherwise it will assume it's a new request | 18:07 |
| sandywalsh | Tushar, the distributed scheduler docs may help http://nova.openstack.org/devref/index.html | 18:08 |
| Tushar | sandywalsh: I will go through this doc to understand more about zone ware scheduler. Thanks. | 18:09 |
| sandywalsh | Tushar, np | 18:09 |
| openstackgerrit | Dolph Mathews proposed a change to openstack/keystone: Moved password check logic to backend https://review.openstack.org/160 | 18:13 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Fix console and debug logging. https://review.openstack.org/161 | 18:14 |
| pvo | vishy: this is the additional state blueprint I said I would get going. Looking for your thoughts or feedback: https://blueprints.launchpad.net/nova/+spec/nova-additional-state' | 18:16 |
| jeblair | dolphm: ping | 18:16 |
| dolphm | jeblair: whats up | 18:16 |
| jeblair | i managed to get a little feedback from jenkins into gerrit about whether a change can be merged | 18:16 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Moved password check logic to backend https://review.openstack.org/160 | 18:17 |
| jeblair | https://review.openstack.org/#change,161 | 18:17 |
| dolphm | jeblair: the fourth line? i just noticed that a few minutes ago :) | 18:17 |
| jeblair | yep | 18:17 |
| dolphm | jeblair: much appreciated! | 18:17 |
| jeblair | so that outputs some explanatory text, so if you see "This change was unable to be automatically merged with the current state of the repository. Please rebase your change and upload a new patchset." you know you can ignore the other failures | 18:18 |
| jeblair | the string formatting is.. unfortunate... but that is, strangely, not a quick fix | 18:18 |
| jeblair | mtaylor: ^ | 18:18 |
| mtaylor | jeblair: wow. love how the \n isn't being processed | 18:19 |
| jeblair | yep. i changed it back, so it's not in there anymore | 18:19 |
| dprince | jeblair: Thank you sir. I added sslverify = false to my gitconfig as well and that worked as well. | 18:20 |
| jeblair | gerrit helpfully styles any messages that have trailing whitespace as preformatted. and jenkins puts in training whitespace i can't get rid of | 18:20 |
| jeblair | dprince: you're welcome! | 18:21 |
| dprince | jeblair: Is there any way we could make the refspec show up instead of the 'branch' on the Gerrit 'status:open' report? | 18:21 |
| dprince | jeblair: The branch name currently always says 'master' which is kind of useless. | 18:21 |
| *** jhtran has joined #openstack-dev | 18:22 | |
| vladimir3p | s1rp: re previous glance issue: installed manually the latest version from git and it helped. thanks | 18:22 |
| jeblair | dprince: i don't think it always will be. i expect we will have multiple branches for things like milestones, etc | 18:22 |
| dolphm | mtaylor: jeblair: can we get jay's updates to test execution in today? https://review.openstack.org/#change,154 | 18:22 |
| mtaylor | dolphm: I'd love to ... what do you need from me on my end? | 18:23 |
| mtaylor | dolphm: do you need me to change the jenkins job before you +2 it? | 18:23 |
| *** alekibango has joined #openstack-dev | 18:24 | |
| dolphm | mtaylor: yeah, look at the last failure.. https://jenkins.openstack.org/job/keystone/101/console | 18:24 |
| dprince | jeblair: Okay. Sure. Is there any way to make the report display 'refspec' as well? | 18:24 |
| dolphm | mtaylor: instead of 'cd keystone/test' 'python run_tests.py', we can just do './run_tests.sh' from the root of the project | 18:24 |
| mtaylor | dolphm: cool. I can do that right now if you like? | 18:25 |
| dolphm | mtaylor: and 'export PYTHONPATH=/home/jenkins/workspace/keystone' definitely isn't needed anymore either... although i'm not sure if you need to run 'python setup.py develop' instead | 18:25 |
| dolphm | mtaylor: yeah, that'd be awesome | 18:25 |
| mtaylor | dolphm: ok. i've made the change - you wanna +2 that bad boy and see how it goes? | 18:26 |
| dolphm | mtaylor: doing it now.. | 18:26 |
| dolphm | mtaylor: hopefully there's no new merge conflicts.. | 18:26 |
| mtaylor | :) | 18:26 |
| mtaylor | dolphm: that's the fun with these large-scale patches | 18:26 |
| jeblair | dprince: i don't think that's tunable, so it would probably be a patch | 18:26 |
| mtaylor | seemed to merge properly | 18:27 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Gets Keystone a bit more inline with the way that other OpenStack projects run tests. Basically, adds the standard run_tests.sh script, modifies the run_tests.py script to do the following: https://review.openstack.org/154 | 18:27 |
| dolphm | mtaylor: this patch also increases test coverage | 18:27 |
| dolphm | Start directory is not importable: 'keystone.test' | 18:27 |
| dolphm | try adding 'python setup.py develop' before running tests | 18:27 |
| dolphm | to fake-install keystone | 18:27 |
| mtaylor | dolphm: so - I'd like to avoid that if we can ... (we don't do that on other projects) - lemme poke for a sec | 18:28 |
| *** martine_ has joined #openstack-dev | 18:28 | |
| *** martine has quit IRC | 18:30 | |
| dolphm | well, between path munging (what other projects do, including parts of keystone), manually setting PYTHONPATH, and running a standard setup.py command that's already built into the project's dependencies (setuptools), i'm vastly in favor of setup.py (either "... develop" or "... install") | 18:30 |
| dprince | jeblair/mtaylor: So long term question. If GitHub adds some features for better review tools will we ditch this Gerrit and just move to GitHub? | 18:31 |
| dolphm | dprince: is github planning code review and continuous integration features? | 18:33 |
| dprince | dolphm: Well they have code review right? I'm not sure what you mean about continuous integration features. Does Gerrit have those? I thought that was more of a Jenkins thing. | 18:35 |
| dprince | dolphm: Its really just an approval mechanism right? | 18:35 |
| mtaylor | dprince: I can't really predict the future ... I would prefer to not do more project process changes in the future unless there is a very compelling reason | 18:35 |
| openstackgerrit | A change was merged to openstack/glance: breaking up MAX_ITEM_LIMIT and making the new values configurable https://review.openstack.org/170 | 18:36 |
| mtaylor | dprince: also, as we move forward, we're building more and more automation on top of this system, so there would be another re-tooling/dev cost that would have to be considered | 18:37 |
| dprince | mtaylor: Sure. We all would. To many people GitHub IS Git however. And I feel like we've side stepped into this Gerrit thing. For good reason (due to the lacking features). But if GitHub adds feature will we be open to using them? | 18:37 |
| dprince | mtaylor: Sorry. I'm just realizing how much detective work I'm going to have to do on top of Gerrit to be able to CI branches with something like Smokestack. | 18:38 |
| mtaylor | dprince: well, you could use the event stream like the jenkins plugin uses | 18:38 |
| *** jshepher has quit IRC | 18:38 | |
| *** jshepher has joined #openstack-dev | 18:39 | |
| mtaylor | dprince: but I'd think that we'd need a compelling set of project requirement changes before undertaking the cost of moving - there are a bunch of more automation tooling things that are in the pipeline to be built on top of this system | 18:39 |
| dprince | mtaylor: Which job should I look at to check that out. | 18:39 |
| mtaylor | dprince: it's inside the gerrit plugin (there is very deep gerrit/jenkins integration - which is one of the reasons that it's a really nice choice for us) | 18:40 |
| mtaylor | dprince: but look at https://review.openstack.org/Documentation/cmd-stream-events.html | 18:40 |
| mtaylor | dprince: which is the basis for the jenkins plugiun | 18:40 |
| mtaylor | dprince: of course - you could _also_ work with us to get the smokestack features you want in to jenkins ... ;) | 18:41 |
| jk0 | mtaylor: is jenkins backed up right now? | 18:41 |
| *** zykes- has quit IRC | 18:42 | |
| dprince | mtaylor: So the gerrit event stream is going to catch things after they have already happened. | 18:43 |
| *** zykes- has joined #openstack-dev | 18:43 | |
| dprince | mtaylor: I'm trying to get a hook into something before we commit it into trunk. Currently it seems like using the 'refspecs' on Gerrit is my only option. | 18:44 |
| mtaylor | jk0: we're doing cloud server snapshots - but doing a real backup strategy is next on our todo list | 18:44 |
| jk0 | mtaylor: ah, I mean the merge queue, sorry :) | 18:44 |
| mtaylor | dprince: the gerrit event stream actually sends you events about approved merge requests | 18:45 |
| jk0 | mtaylor: it doesn't seem to be taking any more Approved MPs | 18:45 |
| mtaylor | dprince: so getting things before we commit them from the stream is exactly how jenkins operates | 18:45 |
| mtaylor | dprince: you can catch an event for everything that happens in gerrit - we have jenkins triggering on review approval - but you can totally watch for new review being submitted and take actions on that | 18:46 |
| mtaylor | dprince: (that's actually how the jenkins plugin works by default- we changed it to work post-approve to prevent against arbitrary malicious people submitting code for review that caused machines to run that code | 18:47 |
| mtaylor | jk0: what do you mean? | 18:48 |
| dprince | mtaylor: Thanks man. I'll check out the Jenkins plugin and see how far I can get. | 18:48 |
| jk0 | mtaylor: I have a MP that's been sitting as 'Approved' for an hour, usually jenkins picks them up within a few minutes | 18:49 |
| mtaylor | dprince: cool. also, jeblair has been hacking in it and has some deeper knowledge than I do | 18:49 |
| mtaylor | jk0: link? | 18:49 |
| jk0 | https://code.launchpad.net/~sandy-walsh/nova/v1_1_novaclient/+merge/70765 | 18:49 |
| *** mwhooker_ has joined #openstack-dev | 18:51 | |
| mtaylor | jaypipes: next time you feel like hacking on keystone tests - could you figure out why I'm having to add PYTHONPATH=`pwd` before running the tests? | 18:51 |
| mtaylor | jk0: it's running tests in that branch now - seems like it might be hung running them | 18:52 |
| mtaylor | jk0: https://jenkins.openstack.org/job/nova-tarmac/110881/console | 18:52 |
| mtaylor | jk0: oh wait! we're both just impatient :) | 18:52 |
| jk0 | interesting | 18:52 |
| *** mwhooker_ has quit IRC | 18:52 | |
| mtaylor | jk0: ok. there's a thing... one sec | 18:52 |
| mtaylor | jk0: that MP is missing a commit message | 18:53 |
| jk0 | sandywalsh: ^ | 18:53 |
| *** mwhooker is now known as mwhooker_ | 18:53 | |
| mtaylor | I'm just going to add one | 18:54 |
| jk0 | doh, I should have caught that | 18:54 |
| sandywalsh | dolphm, out of curiosity, why is the token auth v1.0 and username/password v2.0? | 18:54 |
| mtaylor | well - the system should have caught that | 18:54 |
| mtaylor | jk0: no worries - in the new world order this problem goes away | 18:54 |
| jk0 | nice | 18:54 |
| *** mwhooker has joined #openstack-dev | 18:55 | |
| sandywalsh | jk0, ? looks like jenkins is stuck? | 18:55 |
| vladimir3p | folks, auth-related question: taking into accout recent auth changes what might be the best way to get user details & access code using context? Should I just take user_id and based on that access users DB table? | 18:55 |
| jk0 | sandywalsh: there was no commit message in the MP | 18:55 |
| dolphm | sandywalsh: 1.0/1.1 reflect rackspace's auth... 2.0 will be entirely configurable | 18:55 |
| mtaylor | dolphm: export PYTHONPATH=`pwd` is still required - but seems to fix the test run ... I'm going to bug jaypipes to fix it | 18:55 |
| mtaylor | dolphm: re-running | 18:55 |
| mtaylor | sandywalsh: no worries - I believe I've fixed it | 18:56 |
| dolphm | mtaylor: he also wrote some code to build a virtualenv that may solve the issue as well, but i didn't play with it | 18:56 |
| mtaylor | dolphm: well, we _also_ don't run the jenkins tests in virtualenv :) | 18:56 |
| jaypipes | dolphm: well, we don't run run_tests.sh -V on the jenkins boxess.. | 18:56 |
| sandywalsh | jk0, ah, that causes it to stick? | 18:56 |
| mtaylor | dolphm: I'm basically the hardass who exists to make your life difficult :) | 18:56 |
| sandywalsh | dolphm, gotcha | 18:56 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Gets Keystone a bit more inline with the way that other OpenStack projects run tests. Basically, adds the standard run_tests.sh script, modifies the run_tests.py script to do the following: https://review.openstack.org/154 | 18:56 |
| dolphm | mtaylor: i wouldn't expect you to use a virtualenv | 18:56 |
| mtaylor | DAMMITY | 18:57 |
| dolphm | lol | 18:57 |
| dolphm | hmm - those are test failures that weren't there before | 18:57 |
| mtaylor | jaypipes, dolphm: these seem to be different (and real) errors: https://jenkins.openstack.org/job/keystone/107/console | 18:57 |
| mtaylor | or, when I say real ... actually - ran those commands by hand on that box and they succeeded... hrm | 18:58 |
| mtaylor | dolphm, jaypipes: is it possible that a keystone server from a different run is sneaking in here? | 18:59 |
| dolphm | mtaylor: what do you mean? | 18:59 |
| jaypipes | mtaylor: most definitely it is, because the tests all run the keystone server on the same port :( | 19:00 |
| mtaylor | jaypipes: LOVELY | 19:00 |
| mtaylor | jaypipes: should I limit us to one build thread on that box until we fix that? | 19:00 |
| jaypipes | mtaylor: yes, will have to. | 19:00 |
| dolphm | mtaylor: oh. lol | 19:00 |
| mtaylor | jaypipes: there is a jenkins feature where we can have jenkins create and pass in port numbers ... if that winds up being a helpful thing | 19:01 |
| jaypipes | mtaylor: end goal is to do a little magic like we have in glance's functional test suite that starts up the servers on unused ports in an isolated test dir.. | 19:01 |
| mtaylor | jaypipes: oh - that's even better | 19:01 |
| dolphm | jaypipes: that should be possible soon | 19:01 |
| mtaylor | I limited us to one executor and re-triggered | 19:02 |
| dolphm | grab an unused port from python, throw it into your conf template pattern, and off it goes | 19:02 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Gets Keystone a bit more inline with the way that other OpenStack projects run tests. Basically, adds the standard run_tests.sh script, modifies the run_tests.py script to do the following: https://review.openstack.org/154 | 19:03 |
| dolphm | ldap tests again | 19:03 |
| dolphm | the first test run succeeds - against a sql-only backend | 19:04 |
| *** dprince has quit IRC | 19:04 | |
| *** amccabe has quit IRC | 19:07 | |
| *** deshantm_laptop has joined #openstack-dev | 19:08 | |
| openstackjenkins | Project nova build #1,215: SUCCESS in 3 min 19 sec: http://jenkins.openstack.org/job/nova/1215/ | 19:10 |
| openstackjenkins | Tarmac: Adds OS API 1.1 support | 19:10 |
| dolphm | mtaylor: i blame one of the last two commits (both ldap related) + expanded test coverage in this commit for the test failures... is it possible to totally undo the last two commits in openstack/keystone and then apply this one first? | 19:14 |
| dolphm | (just verified this theory by running tests before/after rebasing) | 19:16 |
| *** jshepher has quit IRC | 19:17 | |
| jeblair | dolphm: maybe the right thing to do would be to git revert the previous commits? (you can revert the commit, which will make a new commit, which you can push to gerrit for review.) | 19:18 |
| dolphm | jeblair: can we then reopen the two closed reviews? | 19:18 |
| jeblair | dolphm: i think you'd have to propose new changes; you can re-open abandoned reviews, but i don't think you can do that for merged ones. | 19:19 |
| jeblair | dolphm: wait | 19:20 |
| dolphm | jeblair: ? | 19:20 |
| jeblair | "Add a "revert change"-button to a submitted patchset | 19:20 |
| jeblair | Clicking "Revert Change" creates a new change with the inverse of the submitted patch set ready for review and submission. This makes it easy to undo a build-breaking change right from the web UI. | 19:20 |
| jeblair | dolphm: so, um, i think that's a button that does exactly what you originally wanted. :) | 19:21 |
| openstackjenkins | Project nova build #1,216: SUCCESS in 3 min 19 sec: http://jenkins.openstack.org/job/nova/1216/ | 19:21 |
| openstackjenkins | Tarmac: This branch adds additional capability to the hosts API extension. The new options allow an admin to reboot or shutdown a host. I also added code to hide this extension if the --allow-admin-api is False, as regular users should have no access to host API calls. | 19:21 |
| jeblair | dolphm: i hadn't noticed it before. :) | 19:21 |
| dolphm | ah, i see that -- let me figure out specifically which revision it was | 19:21 |
| dolphm | oh - only one has made it in lol | 19:21 |
| *** mwhooker has quit IRC | 19:21 | |
| *** amccabe has joined #openstack-dev | 19:22 | |
| mwhooker_ | hi all. I'm having a problem with mox -- I'm wondering if anyone's experienced the same or knows how to fix it. I have a test that's failing on mox.VerifyAll(), but the output of the failure is what seems to be a normal message output 1 character per line, so I have about 30k lines that look like "32097. i" (each line is numbered, for some reason) | 19:23 |
| mtaylor | jeblair: wow. that's a kick-ass feature! | 19:23 |
| *** mwhooker_ is now known as mwhooker | 19:23 | |
| openstackgerrit | Dolph Mathews proposed a change to openstack/keystone: Revert "Add proper simple_bind_s to fakeldap, removed all imports from ldap." https://review.openstack.org/178 | 19:23 |
| mtaylor | dolphm: I may need to re-change the jenkins job to get the inserve patch through | 19:23 |
| mtaylor | one sec | 19:23 |
| mtaylor | dolphm: ok. done | 19:24 |
| dolphm | mtaylor: can i approve the revert? | 19:24 |
| dolphm | ( https://review.openstack.org/#change,178 ) | 19:24 |
| mtaylor | dolphm: yes! | 19:25 |
| *** lorin1 has left #openstack-dev | 19:26 | |
| dolphm | it's about to run the tests that were just failing.. | 19:26 |
| openstackgerrit | A change was merged to openstack/keystone: Revert "Add proper simple_bind_s to fakeldap, removed all imports from ldap." https://review.openstack.org/178 | 19:27 |
| dolphm | mtaylor: now, retrigger https://review.openstack.org/#change,154 ? | 19:28 |
| mtaylor | dolphm: lemme change the jenkins config back | 19:28 |
| openstackgerrit | A change was merged to openstack/keystone: Gets Keystone a bit more inline with the way that other OpenStack projects run tests. Basically, adds the standard run_tests.sh script, modifies the run_tests.py script to do the following: https://review.openstack.org/154 | 19:31 |
| dolphm | yay | 19:31 |
| mtaylor | dolphm: ^^^ yay! | 19:31 |
| dolphm | mtaylor: THANK YOU! :d | 19:31 |
| dolphm | :D * lol | 19:32 |
| jaypipes | dolphm, mtaylor: \o/ | 19:36 |
| openstackgerrit | Monty Taylor proposed a change to openstack/keystone: Moved run_tests.py to match other projects. https://review.openstack.org/179 | 19:37 |
| mtaylor | jaypipes, dolphm: ^^ that removes the need for the PYTHONPATH setting | 19:37 |
| jaypipes | mtaylor: rock. | 19:37 |
| ironcamel | mtaylor: i have a question about tarmac. i checked out the openstack-ci project, and i see the tarmac.conf | 19:37 |
| mtaylor | ironcamel: yes | 19:38 |
| ironcamel | it calls test_nova.sh, which runs bash run_tests.sh -N && python setup.py sdist | 19:38 |
| ironcamel | my question is what is run_tests.sh running against? | 19:38 |
| ironcamel | where is it getting the code from? | 19:38 |
| ironcamel | does it do a bzr clone lp:blah | 19:38 |
| mtaylor | ironcame12: it's running against trunk with the branch in question merged in | 19:38 |
| mtaylor | ironcame12: it does bzr branch lp:project && bzr merge lp:~somebody/project/branch | 19:39 |
| mtaylor | (essentially - it doesn't _ACTUALLY_ run those shell commands :) ) | 19:39 |
| mtaylor | damn. | 19:39 |
| mtaylor | ironcamel: ^^ | 19:39 |
| mtaylor | (stupid tab completion) | 19:39 |
| mtaylor | dolphm: I lied. doesn't work. | 19:39 |
| ironcamel | i am having trouble understanding why tarmac is having problems with https://code.launchpad.net/~rackspace-titan/nova/extensions-osapi1.1-serx/+merge/69749 | 19:39 |
| dolphm | mtaylor: shit lol | 19:40 |
| ironcamel | all the tests pass for me locally | 19:40 |
| mtaylor | dolphm: I'll keep poking | 19:40 |
| mtaylor | ironcame1: looking ... could be deps installed with different versions on the box? | 19:40 |
| ironcamel | and for others as well | 19:40 |
| ironcamel | mtaylor: i have an idea why | 19:40 |
| dolphm | mtaylor: can you cancel a build in progress? lol | 19:40 |
| mtaylor | oh yeah? | 19:40 |
| ironcamel | i have .rng (xml schema files) in the source tree | 19:40 |
| mtaylor | dolphm: I can :) | 19:40 |
| dolphm | mtaylor: well, your change is building | 19:41 |
| mtaylor | dolphm: it won't BREAK anything- it just doesn't actually fix it all the way | 19:41 |
| jaypipes | ironcamel: XML? Well there's your problem! ;P | 19:41 |
| ironcamel | mtaylor: it seems as though tarmac somehow is not copying those files over | 19:41 |
| mtaylor | ironcame1: hrm. oh - wait - is it possible that you didn't bzr add them? | 19:41 |
| mtaylor | oh - no, I see them there | 19:42 |
| ironcamel | jaypipes: hehe. i remember from the conference you were not a fan of xml. neither am i really. just trying to get this branch in. | 19:42 |
| jaypipes | :) | 19:42 |
| jaypipes | ironcamel: I know, was just kiddin you | 19:42 |
| mtaylor | ironcame1: actually ... /tmp/tmp_11D8R/nova/__init__.py/nova/api/openstack/schemas/v1.1/extension.rng | 19:43 |
| * jk0 remembers that XML talk | 19:43 | |
| mtaylor | ironcame1: that seems like a very odd file path | 19:43 |
| jk0 | that was my favorite talk of the summit | 19:43 |
| *** ameade has quit IRC | 19:43 | |
| ironcamel | yeah, why is it constructing it like that ... hmmm | 19:44 |
| ironcamel | http://paste.openstack.org/show/2123/ | 19:45 |
| ironcamel | that is the code that constructs the path to the xml schema file | 19:45 |
| ironcamel | is there a bug in that? | 19:45 |
| mtaylor | ironcame1: so - the essential sequence is something like "TEMPDIR=mktempdir ; bzr branch lp:nova $TEMPDIR; cd $TEMPDIR ; bzr merge lp:~rackspace-titan/nova/extensions-osapi1.1-serx ; ./run_tests.sh" | 19:46 |
| ironcamel | utils.novadir() must be returning /tmp/tmp_11D8R/nova/__init__.py | 19:46 |
| mtaylor | ironcame1: it's possible that utils.novadir() is doing something weird | 19:46 |
| mtaylor | yeah | 19:46 |
| mtaylor | return os.path.abspath(nova.__file__).split('nova/__init__.pyc')[0] | 19:47 |
| *** ameade has joined #openstack-dev | 19:47 | |
| ironcamel | maybe it should not be splitting on .pyc | 19:47 |
| ironcamel | should it be os.path.abspath(nova.__file__).split('nova/__init__.py')[0] ? | 19:48 |
| mtaylor | that works WAY better for me | 19:48 |
| mtaylor | although I suppose sometimes it might return __init__.py | 19:48 |
| ironcamel | so i'll try changing that function, and pushing my branch again | 19:49 |
| ironcamel | does that sound ok? | 19:49 |
| mtaylor | ironcame1: I think you might break other things- I think the function needs to account for nova.__file__ possibly equalling __init__.py and __init__.pyc | 19:49 |
| ironcamel | ok | 19:50 |
| mtaylor | ironcamel: it's ugly - but os.path.split(os.path.split(os.path.abspath(nova.__file__))[0])[0] should work for both cases | 19:50 |
| ironcamel | so you think the first time .__file__ is the .py file, and each subsequent time it is the .pyc file? | 19:51 |
| mtaylor | ironcamel: I suck ... os.path.abspath(nova.__file__).split(nova.__file__)[0] | 19:52 |
| mtaylor | ironcamel: would be the, you know, consistent thing :) | 19:52 |
| ironcamel | much better :) | 19:54 |
| openstackjenkins | Project nova build #1,217: SUCCESS in 3 min 23 sec: http://jenkins.openstack.org/job/nova/1217/ | 19:55 |
| openstackjenkins | Tarmac: nova.exception.wrap_exception will re-raise some exceptions, but in the process of possibly notifying that an exception has occurred, it may clobber the current exception information. nova.utils.to_primitive in particular (used by the notifier code) will catch and handle an exception clobbering the current exception being handled in wrap_exception. Eventually when using the bare 'raise', it will attempt to rai | 19:55 |
| openstackjenkins | The patch saves the exception at the beginning of wrap_exception and then re-raises the original exception avoiding the possibility of a clobbered exception. | 19:55 |
| kpepple | jaypipes: time for a quick question ? | 20:01 |
| jaypipes | kpepple: of course | 20:01 |
| kpepple | in cactus version of glance, when configured to persist images to Swift, does it store them all under a single user/id ? | 20:02 |
| jaypipes | kpepple: yes. the swift_store_user/key config options. | 20:04 |
| jaypipes | kpepple: that said, you *can* however specify an alternate swift location via the X-Image-Meta-Location property, in which case Glance is instructed to fetch the image from an alternate Swift location... | 20:05 |
| kpepple | jaypipes: hhhmmm ... ok, thought so. second question: are nova and glance versions pegged -- can i use glance diablo with nova cactus ? | 20:06 |
| jaypipes | kpepple: no, you cannot. = | 20:07 |
| jaypipes | kpepple: you must use cactus for both, since there was a backwards incompatible client and API versioning change in Glance D1 | 20:07 |
| jaypipes | kpepple: though we are recommending running D3 and later for both Nova and Glance... | 20:08 |
| kpepple | jaypipes: ok, thought it was a long shot with all the changes ... | 20:08 |
| kpepple | jaypipes: the version horse left the barn several months ago ... | 20:08 |
| jaypipes | kpepple: indeed :) | 20:08 |
| *** dprince has joined #openstack-dev | 20:11 | |
| kpepple | jaypipes: thanks. trying to figure out billing on saved images ... | 20:13 |
| jaypipes | kpepple: billing on saved images? | 20:13 |
| kpepple | jaypipes: we bill on user's image storage | 20:14 |
| jaypipes | I see | 20:15 |
| *** nickon has quit IRC | 20:17 | |
| dprince | mtaylor: Any idea how long it takes the PPA to update novaclient? | 20:18 |
| dprince | I need the latest version bzr70 and its still at bzr65. | 20:18 |
| dprince | https://jenkins.openstack.org/job/python-novaclient-ppa/26/label=nova,series=maverick/console | 20:23 |
| openstackjenkins | Project nova build #1,218: SUCCESS in 3 min 14 sec: http://jenkins.openstack.org/job/nova/1218/ | 20:25 |
| openstackjenkins | Tarmac: Set image progress to 100 if the image is active. | 20:25 |
| dprince | According to that we had success uploading to the PPA. But when I go to install them the package isn't there. Additionally the 'Packages' list in the PPA is stale (still back at revision 65). | 20:25 |
| *** dprince has quit IRC | 20:26 | |
| mtaylor | dprince: it can take a bit - they have a queue | 20:26 |
| mtaylor | well poo | 20:26 |
| openstackjenkins | Project burrow build #26: SUCCESS in 11 sec: http://jenkins.openstack.org/job/burrow/26/ | 20:31 |
| openstackjenkins | Tarmac: Added more unit test coverage around update_messages. | 20:31 |
| *** deshantm_laptop has quit IRC | 20:49 | |
| jaypipes | bcwaldon, blamar_, dprince: w00t. Swift functional test case done in Glance: http://paste.openstack.org/show/2125/ | 20:59 |
| bcwaldon | jaypipes: tested AND working?! | 20:59 |
| jaypipes | bcwaldon: indeed :) | 20:59 |
| bcwaldon | jaypipes: that's awesome, good work! | 20:59 |
| jaypipes | bcwaldon: thx :) | 21:01 |
| openstackgerrit | A change was merged to openstack/glance: Add @skip_if_disabled decorator to test.utils and integrate it into the base functional API test case. The S3 functional test case now uses test_api.TestApi as its base class and the setUp() method sets the disabled and disabled_message attributes that th https://review.openstack.org/177 | 21:06 |
| openstackgerrit | Jay Pipes proposed a change to openstack/glance: Consolidates the functional API test cases into /glance/tests/functional/test_api.py, adds a new Swift functional test case, verified that it works on Cloud Files with a test account. https://review.openstack.org/180 | 21:06 |
| *** bengrue has joined #openstack-dev | 21:06 | |
| jaypipes | mtaylor: once https://review.openstack.org/#change,180 is approved, I'd like to upload to the glance builder slave some S3 and Swift config files with test account creds and ensure we are testing S3 and Swift drivers on each commit... | 21:09 |
| *** martine_ has quit IRC | 21:16 | |
| *** rnirmal has quit IRC | 21:25 | |
| *** cp16net has quit IRC | 21:26 | |
| openstackgerrit | Dolph Mathews proposed a change to openstack/keystone: Refactoring business logic behind GET /tenants to make it less convoluted. https://review.openstack.org/181 | 21:26 |
| *** lts has quit IRC | 21:26 | |
| openstackgerrit | James E. Blair proposed a change to openstack/openstack-ci-puppet: Add iptables module and rules to puppet. https://review.openstack.org/182 | 21:33 |
| *** rnorwood has quit IRC | 21:48 | |
| *** bcwaldon has quit IRC | 22:02 | |
| *** anotherjesse has joined #openstack-dev | 22:18 | |
| anotherjesse | sandywalsh: does the changes we are sending in with pull 56, 65 make sense now? | 22:19 |
| sandywalsh | anotherjesse, yes, thanks ... once we get the redirect stuff straightened out we'll pull those right in | 22:26 |
| sandywalsh | anotherjesse (actually they may not conflict anyway) | 22:26 |
| *** bsza has quit IRC | 22:26 | |
| anotherjesse | does the comment in the redirect about rs auth make sense? | 22:26 |
| anotherjesse | https://github.com/rackspace/python-novaclient/pull/62#issuecomment-1758229 | 22:26 |
| sandywalsh | yes, I think the service catalog is the thing I was missing. Still have to read to figure out how a single service is specified (if no service catalog) | 22:27 |
| sandywalsh | ie. we only have nova and nothing else, so always redirect to nova api after auth | 22:28 |
| sandywalsh | where do I tell keystone that info | 22:28 |
| sandywalsh | degenerate case of service catalog I guess | 22:28 |
| anotherjesse | currently you have to tell keystone a list of endpoints for each tenant (or have a default one for all tenants) | 22:28 |
| anotherjesse | that is an implementation detail though | 22:28 |
| anotherjesse | the API says auth token returns a catalog … how it is generated is not specified ... | 22:29 |
| anotherjesse | rax idm will probably do something specific to your needs? | 22:29 |
| *** Tushar has quit IRC | 22:29 | |
| *** markvoelker has quit IRC | 22:31 | |
| *** kbringard has quit IRC | 22:32 | |
| *** sandywalsh has quit IRC | 22:33 | |
| *** CatKiller has quit IRC | 22:36 | |
| *** zigo has joined #openstack-dev | 22:39 | |
| *** mattray has quit IRC | 22:39 | |
| *** amccabe has quit IRC | 22:44 | |
| *** sandywalsh has joined #openstack-dev | 22:46 | |
| mtaylor | jaypipes: ok. let me know where the files are | 22:47 |
| jaypipes | mtaylor: I will, as soon as other Glance devs approve that proposed branch... | 22:48 |
| openstackgerrit | Yogeshwar Srikrishnan proposed a change to openstack/keystone: Changes to allow additional calls to support endpoint template CRUD and additional checks on existing method. https://review.openstack.org/183 | 22:48 |
| *** alekibango has quit IRC | 22:52 | |
| *** CatKiller has joined #openstack-dev | 22:53 | |
| *** yogirackspace has left #openstack-dev | 22:59 | |
| *** jhtran has quit IRC | 23:13 | |
| *** sandywalsh has quit IRC | 23:25 | |
| *** CatKiller has quit IRC | 23:26 | |
| openstackjenkins | Project nova build #1,219: SUCCESS in 3 min 13 sec: http://jenkins.openstack.org/job/nova/1219/ | 23:29 |
| openstackjenkins | Tarmac: Adds the ability to read/write to a local xenhost config. No changes to the nova codebase; this will be used only by admin tools that have yet to be created. | 23:29 |
| *** sandywalsh has joined #openstack-dev | 23:38 | |
| *** vladimir3p has quit IRC | 23:43 | |
| *** jkoelker has quit IRC | 23:43 | |
| *** mfer has quit IRC | 23:47 | |
| *** mfer has joined #openstack-dev | 23:47 | |
| *** martine_ has joined #openstack-dev | 23:48 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!