| *** metral_ has joined #openstack-dev | 00:02 | |
| *** metral has quit IRC | 00:05 | |
| *** metral_ is now known as metral | 00:06 | |
| *** mrodden1 has quit IRC | 00:06 | |
| *** rnirmal has quit IRC | 00:12 | |
| *** ewindisch has quit IRC | 00:12 | |
| *** cp16net is now known as cp16net|away | 00:14 | |
| *** rpedde is now known as rpedde_away | 00:16 | |
| *** anniec has joined #openstack-dev | 00:19 | |
| *** maxiz has joined #openstack-dev | 00:19 | |
| *** mnewby_ has joined #openstack-dev | 00:21 | |
| *** anniec_ has joined #openstack-dev | 00:22 | |
| *** mnewby_ has quit IRC | 00:23 | |
| *** anniec has quit IRC | 00:24 | |
| *** anniec_ is now known as anniec | 00:24 | |
| *** winston-d has joined #openstack-dev | 00:24 | |
| *** mnewby has quit IRC | 00:25 | |
| *** metral has quit IRC | 00:25 | |
| *** metral has joined #openstack-dev | 00:27 | |
| *** mnewby has joined #openstack-dev | 00:27 | |
| *** mrodden has joined #openstack-dev | 00:28 | |
| *** maurosr has joined #openstack-dev | 00:28 | |
| *** mrodden1 has joined #openstack-dev | 00:29 | |
| *** mnewby_ has joined #openstack-dev | 00:29 | |
| *** clarkb has quit IRC | 00:32 | |
| *** mrodden has quit IRC | 00:32 | |
| *** maurosr has quit IRC | 00:33 | |
| *** mnewby has quit IRC | 00:33 | |
| *** mnewby_ is now known as mnewby | 00:33 | |
| *** sniperd has quit IRC | 00:36 | |
| *** maurosr has joined #openstack-dev | 00:37 | |
| *** epim_ has joined #openstack-dev | 00:37 | |
| *** hemna has quit IRC | 00:38 | |
| *** epim has quit IRC | 00:40 | |
| *** epim_ is now known as epim | 00:40 | |
| *** metral has quit IRC | 00:40 | |
| *** maurosr has quit IRC | 00:41 | |
| *** dims has quit IRC | 00:41 | |
| *** metral has joined #openstack-dev | 00:41 | |
| alexpilotti | mikal: hi! | 00:41 |
|---|---|---|
| mikal | alexpilotti: hi | 00:43 |
| alexpilotti | mikal: I'm porting config drive v2 | 00:43 |
| mikal | Ahhh, right. That email thread from yesterday. | 00:43 |
| alexpilotti | mikal: there's a bug in nova\api\metadata\base.py | 00:43 |
| alexpilotti | mikal: os.path.normpath(path) :-) | 00:44 |
| mikal | Hmmm, you should fix that! | 00:44 |
| alexpilotti | mikal: sure, np. The question is: we found a lot of places where URLs are normalized with os.path.normpath, which replaces "/" with "\" on Windows :-) | 00:45 |
| mikal | Hmmmm | 00:46 |
| mikal | So there are two things here | 00:46 |
| mikal | Config drive, which writes ISO etc | 00:46 |
| *** alunduil has joined #openstack-dev | 00:46 | |
| mikal | And then the metadata server, which is another way of getting to the data | 00:46 |
| *** esp2 has left #openstack-dev | 00:46 | |
| mikal | The config drive hooks into metadata server to reduce code duplication | 00:46 |
| alexpilotti | mikal: this is code called by config drive | 00:46 |
| mikal | But I didn't write much of the metadata server | 00:46 |
| mikal | Yep, agreed | 00:47 |
| alexpilotti | mikal: in: def metadata_for_config_drive(self): | 00:47 |
| mikal | But I guess I'm saying I'd be surprised if anyone has tried to run the metadata server on windows | 00:47 |
| mikal | And it might require cleanup | 00:47 |
| mikal | Sounds like you should file a bug and either fix it or I can fix it | 00:47 |
| mikal | But most of this code was written by smoser IIRC | 00:47 |
| alexpilotti | mikal: sure, the question is: is there a way to warn about using os.path.normpath to normalize URLs? | 00:47 |
| mikal | Hmmmm, that's hard right? | 00:48 |
| alexpilotti | mikal: beside metadata, there's a LOT of places where we found this way of normalizing URLs | 00:48 |
| mikal | How does the code know that its going to be used for a URL? | 00:48 |
| *** sacharya1 has joined #openstack-dev | 00:48 | |
| mikal | It sounds like a lint check, but I don't know how to express it | 00:48 |
| alexpilotti | mikal: the code not, but a dev should know when he's normalizing urls or paths | 00:48 |
| mikal | True | 00:49 |
| mikal | I guess I was going for an automated check here | 00:49 |
| mikal | Otherwise you're going to be fixing new occurrances of this for the rest of your life | 00:49 |
| alexpilotti | mikal: one option could be to monkey patch os.path.normpath suring the tests | 00:49 |
| *** matiu_ is now known as matiu | 00:49 | |
| mikal | Or add unit tests to stop regressions every time you find an example... | 00:50 |
| mikal | But that sounds laborious | 00:50 |
| alexpilotti | mikal: other option is to have jenkins running on Windows as well | 00:50 |
| *** sacharya has quit IRC | 00:50 | |
| alexpilotti | mikal: in that case the unit tests should fail | 00:50 |
| mikal | Hmmm, that sounds like something we should talk to mordred about | 00:51 |
| alexpilotti | mikal: cool | 00:51 |
| mikal | If not jenkins, a smokestack style test | 00:51 |
| mikal | Which would be a very good idea | 00:51 |
| mikal | Especially if it could run hyper V tests against a real hyper V install instead of mocks | 00:51 |
| alexpilotti | mikal: we are working with Microsoft to set up smokestack | 00:51 |
| mikal | Yep. I think that's really important to get done. | 00:51 |
| *** clarkb has joined #openstack-dev | 00:51 | |
| alexpilotti | mikal: cool, I'll get back in touch as soon as we start working on that | 00:52 |
| alexpilotti | mikal: tx | 00:52 |
| mikal | Sure, NP | 00:52 |
| alexpilotti | mikal: changing subject: | 00:52 |
| alexpilotti | mikal: the tool you used to generate the isos is genisoimage | 00:53 |
| *** salv-orlando has quit IRC | 00:53 | |
| *** maurosr has joined #openstack-dev | 00:53 | |
| alexpilotti | mikal: we are using a port of mkisofs on Windows (the later non GPL one, compiled with mingw to avoid licensing issues) | 00:54 |
| mikal | Yep, I saw you wanted to do that | 00:54 |
| mikal | I'm fine with that | 00:54 |
| vishy | Q310: cool | 00:54 |
| mikal | Perhaps add a flag which controls what is used? | 00:54 |
| vishy | Q310: feel free to propose it | 00:54 |
| alexpilotti | mikal: that was exactly where I was heading ;-) | 00:54 |
| *** dims has joined #openstack-dev | 00:54 | |
| alexpilotti | mikal: is it ok for you to specify in the flag the name / path of the command? | 00:55 |
| mikal | Hmmm. | 00:55 |
| * mikal ponders | 00:55 | |
| mikal | How do people install this stuff on windows? | 00:55 |
| mikal | Will the command always be in the same place because it is installed by the nova installer or something? | 00:55 |
| alexpilotti | mikal: with our installer: http://www.cloudbase.it/openstack/openstack-compute-installer/ | 00:56 |
| alexpilotti | mikal: or manually of course if they prefer | 00:56 |
| alexpilotti | mikal: we generate nova.conf suring the install process, so the user doesn't have to take care of anything | 00:57 |
| mikal | Hmmm, so the binary should be in a known location right? | 00:57 |
| alexpilotti | mikal: what I'd like to avoid, is to put mkisofs in th epath | 00:57 |
| mikal | I guess I'd prefer not to pass an entire path | 00:57 |
| mikal | But I don't have a strong reason for why I prefer that | 00:57 |
| mikal | Well, if the flag was --use_mkisofs then the code could assume a full path? | 00:57 |
| alexpilotti | mikal: by passing the name only, it will look for in in the PATH | 00:57 |
| *** maurosr has quit IRC | 00:57 | |
| *** mnewby has quit IRC | 00:58 | |
| alexpilotti | mikal: so if we name it eg: mkfsiso_cmd | 00:58 |
| mikal | Sure, ok | 00:58 |
| mikal | I think that is fine | 00:58 |
| alexpilotti | mikal: great, no more questions so far :-) tx! | 00:59 |
| *** lloydde_ has quit IRC | 01:00 | |
| *** maurosr has joined #openstack-dev | 01:00 | |
| mikal | NP | 01:01 |
| *** dolphm has joined #openstack-dev | 01:01 | |
| *** TerryH has quit IRC | 01:03 | |
| *** maurosr has quit IRC | 01:05 | |
| *** datsun180b has quit IRC | 01:08 | |
| *** sunilthaha has joined #openstack-dev | 01:08 | |
| *** sunilthaha has joined #openstack-dev | 01:09 | |
| *** sunilthaha has quit IRC | 01:10 | |
| *** rkukura has quit IRC | 01:12 | |
| *** mrodden1 has quit IRC | 01:12 | |
| *** gugl has quit IRC | 01:12 | |
| *** dprince has quit IRC | 01:13 | |
| *** danwent has joined #openstack-dev | 01:14 | |
| *** rkukura has joined #openstack-dev | 01:14 | |
| *** sunilthaha has joined #openstack-dev | 01:14 | |
| *** maurosr has joined #openstack-dev | 01:17 | |
| *** wenjianhn has joined #openstack-dev | 01:19 | |
| *** e1mer has quit IRC | 01:20 | |
| *** srn has joined #openstack-dev | 01:21 | |
| *** jog0 has quit IRC | 01:21 | |
| *** maurosr has quit IRC | 01:21 | |
| *** srn has quit IRC | 01:21 | |
| *** patelna__ has quit IRC | 01:22 | |
| *** e1mer has joined #openstack-dev | 01:24 | |
| *** lloydde has joined #openstack-dev | 01:25 | |
| *** rnirmal has joined #openstack-dev | 01:31 | |
| *** maurosr has joined #openstack-dev | 01:34 | |
| *** vipul has quit IRC | 01:35 | |
| *** dprince has joined #openstack-dev | 01:36 | |
| *** timello has joined #openstack-dev | 01:37 | |
| *** maurosr has quit IRC | 01:39 | |
| *** mnewby has joined #openstack-dev | 01:40 | |
| *** jcmartin has quit IRC | 01:40 | |
| *** slackguru has joined #openstack-dev | 01:42 | |
| *** Q310 has quit IRC | 01:42 | |
| *** mnewby has quit IRC | 01:43 | |
| *** littleidea has quit IRC | 01:44 | |
| *** slackguru has left #openstack-dev | 01:44 | |
| *** xchu has joined #openstack-dev | 01:46 | |
| *** cp16net|away is now known as cp16net | 01:48 | |
| *** s0mik has quit IRC | 01:48 | |
| *** maurosr has joined #openstack-dev | 01:52 | |
| *** metral_ has joined #openstack-dev | 01:53 | |
| *** dprince has quit IRC | 01:54 | |
| *** metral has quit IRC | 01:55 | |
| *** metral_ is now known as metral | 01:55 | |
| *** maurosr has quit IRC | 01:57 | |
| *** maurosr has joined #openstack-dev | 01:58 | |
| *** jaypipes has quit IRC | 01:59 | |
| *** esp1 has joined #openstack-dev | 01:59 | |
| *** SlickNik has quit IRC | 02:00 | |
| *** troytoman-away has quit IRC | 02:03 | |
| *** ayoung-afk is now known as ayoung | 02:03 | |
| ayoung | dolphm, when you get the chance: https://review.openstack.org/#/c/15429/ and https://review.openstack.org/#/c/15543/ | 02:03 |
| *** maurosr has quit IRC | 02:03 | |
| *** Apsu` has quit IRC | 02:03 | |
| dolphm | ayoung: sure | 02:04 |
| *** networkstatic has joined #openstack-dev | 02:04 | |
| ayoung | dolphm, also, let me know if any of yours need review. AFAICT I have no reviews in my queue | 02:04 |
| dolphm | ayoung: i put up one for v3 policies on keystone master this morning | 02:04 |
| dolphm | ayoung: same review as what was pending on the feature branch, but cut out the endpoint_id association | 02:05 |
| *** metral has quit IRC | 02:05 | |
| ayoung | dolphm, looking. Ialso see the extra-twice one. I'll see if there is anything else on the watched list | 02:05 |
| dolphm | ayoung: let me finish some v3 catalog stuff (working on your feedback from way back on that lol) before i poke at your reviews, i want to test them both on my machine | 02:06 |
| *** eglynn_ has joined #openstack-dev | 02:06 | |
| *** troytoman-away has joined #openstack-dev | 02:07 | |
| ayoung | dolphm, why user_ref['extra'].pop('tenants', None) | 02:07 |
| *** lloydde_ has joined #openstack-dev | 02:07 | |
| ayoung | dolphm, also, adding dkranz as reviewer on that one | 02:08 |
| *** eglynn has quit IRC | 02:08 | |
| *** Apsu` has joined #openstack-dev | 02:09 | |
| dolphm | ayoung: tenant memberships used to be shoved into the 'extra' column, if they're still there, they shouldn't be returned | 02:09 |
| *** littleidea has joined #openstack-dev | 02:09 | |
| dolphm | ayoung: coding for paranoia | 02:09 |
| *** lloydde has quit IRC | 02:10 | |
| *** metral has joined #openstack-dev | 02:10 | |
| *** roadTripper has quit IRC | 02:10 | |
| *** tgall_foo has joined #openstack-dev | 02:11 | |
| *** tgall_foo has quit IRC | 02:11 | |
| *** tgall_foo has joined #openstack-dev | 02:11 | |
| *** darjeeling has quit IRC | 02:13 | |
| *** sunilthaha has quit IRC | 02:16 | |
| *** metral has quit IRC | 02:16 | |
| *** cp16net is now known as cp16net|away | 02:19 | |
| *** simonmcc1 has quit IRC | 02:19 | |
| *** metral has joined #openstack-dev | 02:19 | |
| *** lloydde_ has quit IRC | 02:19 | |
| *** cp16net|away is now known as cp16net | 02:21 | |
| *** littleidea has quit IRC | 02:22 | |
| metral | questions: can i use the set metadata to change the image_type to be something other than "base" or "server" or is that hardcoded by openstack | 02:22 |
| *** Mandell_ has quit IRC | 02:22 | |
| *** mnewby has joined #openstack-dev | 02:25 | |
| *** littleidea has joined #openstack-dev | 02:27 | |
| *** sunilthaha has joined #openstack-dev | 02:29 | |
| *** metral has quit IRC | 02:31 | |
| *** timello has quit IRC | 02:33 | |
| *** rkukura has quit IRC | 02:35 | |
| *** tgall_foo has quit IRC | 02:36 | |
| *** epim_ has joined #openstack-dev | 02:42 | |
| *** epim has quit IRC | 02:43 | |
| *** epim_ is now known as epim | 02:43 | |
| ayoung | dolphm, also https://review.openstack.org/#/c/15403/ THat is the authenticate refactoring. It is not a 100% solution, but it is a step in the right direction, and we should finish this before tackiling the ec2 token issue | 02:43 |
| *** novas0x2a|laptop has quit IRC | 02:43 | |
| dolphm | ayoung: ec2 token issue? | 02:43 |
| *** epim has quit IRC | 02:44 | |
| *** tgall_foo has joined #openstack-dev | 02:47 | |
| *** tgall_foo has quit IRC | 02:47 | |
| *** tgall_foo has joined #openstack-dev | 02:47 | |
| *** zbitter has joined #openstack-dev | 02:48 | |
| ayoung | dolphm, https://bugs.launchpad.net/keystone/+bug/1064914 | 02:49 |
| uvirtbot` | Launchpad bug 1064914 in keystone "Able to access ec2 resources with out a user-role" [Critical,Triaged] | 02:49 |
| *** zaneb has quit IRC | 02:51 | |
| *** amotoki has joined #openstack-dev | 02:54 | |
| dolphm | ayoung: i wasn't aware of that bug ... do you know what's causing it? | 02:55 |
| ayoung | dolphm, yeah. The authenticate method is not called in that case. The code executed is under contrib. heckj did a quick look at it and confirmed as well. | 02:59 |
| openstackgerrit | A change was merged to openstack/nova: Decouple EC2 API from using instance id https://review.openstack.org/15391 | 02:59 |
| dolphm | ayoung: no token validation? | 03:00 |
| ayoung | dolphm, https://github.com/openstack/keystone/blob/master/keystone/contrib/ec2/core.py#L148 | 03:00 |
| ayoung | token with no roles, I think | 03:00 |
| lifeless | /win 52 | 03:00 |
| ayoung | dolphm, so the authenticate refactoring happens first, and then we use common code for that and for all of the contrib tokens authenticates | 03:01 |
| dolphm | ayoung: wow, i didn't know this code was even here | 03:02 |
| *** dims has quit IRC | 03:02 | |
| ayoung | dolphm, there are a couple in contrib | 03:04 |
| dolphm | ayoung: :( | 03:04 |
| *** markmcclain has joined #openstack-dev | 03:05 | |
| *** melwitt has quit IRC | 03:10 | |
| *** jimp has joined #openstack-dev | 03:12 | |
| *** anniec has quit IRC | 03:15 | |
| *** jcooley has joined #openstack-dev | 03:23 | |
| *** sniperd has joined #openstack-dev | 03:29 | |
| openstackgerrit | A change was merged to openstack/swift: Fixed swift-bench when using -b and a conf file https://review.openstack.org/15603 | 03:32 |
| *** danwent has quit IRC | 03:35 | |
| *** PaulM has left #openstack-dev | 03:35 | |
| openstackgerrit | A change was merged to openstack/keystone: HACKING compliance: consistent use of 'except' https://review.openstack.org/15417 | 03:36 |
| *** ayoung has quit IRC | 03:38 | |
| *** jimp has quit IRC | 03:41 | |
| *** gyee has joined #openstack-dev | 03:44 | |
| *** alexpilotti has quit IRC | 03:48 | |
| *** jcooley has quit IRC | 03:53 | |
| *** zhuadl has joined #openstack-dev | 03:54 | |
| *** mnewby has quit IRC | 03:57 | |
| openstackgerrit | A change was merged to openstack/keystone: monkeypatch cms Popen https://review.openstack.org/15429 | 03:59 |
| openstackgerrit | A change was merged to openstack/keystone: SQL upgrade test. https://review.openstack.org/15543 | 03:59 |
| *** jakedahn_zz is now known as jakedahn | 04:03 | |
| *** rkukura has joined #openstack-dev | 04:04 | |
| *** anniec72 has joined #openstack-dev | 04:05 | |
| *** anniec72 has quit IRC | 04:05 | |
| *** aeperezt has quit IRC | 04:07 | |
| *** e1mer has quit IRC | 04:10 | |
| *** e1mer has joined #openstack-dev | 04:11 | |
| *** winston-d has quit IRC | 04:19 | |
| *** sacharya has joined #openstack-dev | 04:21 | |
| *** sacharya1 has quit IRC | 04:23 | |
| *** danwent has joined #openstack-dev | 04:24 | |
| *** vipul_ has joined #openstack-dev | 04:28 | |
| openstackgerrit | A change was merged to openstack/keystone: Fixed typo in log message https://review.openstack.org/14827 | 04:37 |
| *** arata has joined #openstack-dev | 04:40 | |
| *** Mandell has joined #openstack-dev | 04:42 | |
| *** mnewby has joined #openstack-dev | 04:46 | |
| *** mnewby has quit IRC | 04:50 | |
| openstackgerrit | A change was merged to openstack/nova: Corrects usage of db.api.network_get https://review.openstack.org/15434 | 04:56 |
| *** diganta has joined #openstack-dev | 04:59 | |
| *** sacharya has quit IRC | 05:00 | |
| *** adjohn has quit IRC | 05:01 | |
| *** adjohn has joined #openstack-dev | 05:01 | |
| diganta | hi good morning all | 05:02 |
| *** markmcclain has quit IRC | 05:04 | |
| *** zaitcev has quit IRC | 05:05 | |
| *** jcooley has joined #openstack-dev | 05:05 | |
| *** zhuadl has quit IRC | 05:18 | |
| *** s0mik has joined #openstack-dev | 05:18 | |
| *** e1mer has quit IRC | 05:19 | |
| *** e1mer has joined #openstack-dev | 05:20 | |
| *** mikal has quit IRC | 05:22 | |
| *** vipul_ has quit IRC | 05:23 | |
| *** vipul has joined #openstack-dev | 05:24 | |
| *** jcooley has quit IRC | 05:25 | |
| *** mikal has joined #openstack-dev | 05:28 | |
| *** garyk has quit IRC | 05:31 | |
| *** wenjianhn has quit IRC | 05:31 | |
| *** dolphm has quit IRC | 05:37 | |
| *** wenjianhn has joined #openstack-dev | 05:38 | |
| *** e1mer has quit IRC | 05:41 | |
| *** jcooley has joined #openstack-dev | 05:41 | |
| *** nati_ueno has joined #openstack-dev | 05:43 | |
| *** lloydde has joined #openstack-dev | 05:43 | |
| *** e1mer has joined #openstack-dev | 05:46 | |
| *** jcooley has quit IRC | 05:49 | |
| *** jcooley has joined #openstack-dev | 05:49 | |
| *** darjeeling has joined #openstack-dev | 05:50 | |
| *** gyee has quit IRC | 05:50 | |
| *** sunilthaha has quit IRC | 05:51 | |
| *** vipuls has joined #openstack-dev | 05:55 | |
| *** vipul has quit IRC | 05:59 | |
| *** hattwick has quit IRC | 05:59 | |
| *** tgall_foo has quit IRC | 05:59 | |
| *** adjohn has quit IRC | 05:59 | |
| *** nati_ueno has quit IRC | 06:04 | |
| *** vipuls has quit IRC | 06:04 | |
| *** nati_ueno has joined #openstack-dev | 06:04 | |
| *** vipul has joined #openstack-dev | 06:05 | |
| *** Mandell has quit IRC | 06:05 | |
| *** vipul is now known as vipul|away | 06:07 | |
| *** vipul|away is now known as vipul | 06:07 | |
| *** nati_ueno_2 has joined #openstack-dev | 06:09 | |
| *** nati_ueno has quit IRC | 06:10 | |
| *** joearnold has quit IRC | 06:10 | |
| *** anderstj has quit IRC | 06:10 | |
| *** anderstj has joined #openstack-dev | 06:11 | |
| *** annegentle has quit IRC | 06:12 | |
| *** annegentf has joined #openstack-dev | 06:13 | |
| *** joearnold has joined #openstack-dev | 06:14 | |
| *** markwash has quit IRC | 06:14 | |
| *** Mandell has joined #openstack-dev | 06:15 | |
| *** Mandell has quit IRC | 06:18 | |
| *** Shrews has quit IRC | 06:24 | |
| *** Shrews has joined #openstack-dev | 06:26 | |
| *** Mandell has joined #openstack-dev | 06:27 | |
| *** jtran has quit IRC | 06:28 | |
| *** markwash has joined #openstack-dev | 06:30 | |
| *** jtran has joined #openstack-dev | 06:31 | |
| *** jtran has quit IRC | 06:34 | |
| *** nati_ueno_2 has quit IRC | 06:34 | |
| *** nati_ueno has joined #openstack-dev | 06:34 | |
| *** nati_ueno_2 has joined #openstack-dev | 06:37 | |
| *** nati_ueno has quit IRC | 06:37 | |
| *** rushiagr has joined #openstack-dev | 06:40 | |
| *** hattwick has joined #openstack-dev | 06:42 | |
| *** cmagina has quit IRC | 06:45 | |
| *** sunilthaha has joined #openstack-dev | 06:45 | |
| *** mrunge has joined #openstack-dev | 06:46 | |
| *** lloydde has quit IRC | 06:46 | |
| *** Mandell has quit IRC | 06:48 | |
| *** gongysh has quit IRC | 06:49 | |
| *** garyk has joined #openstack-dev | 06:50 | |
| openstackgerrit | A change was merged to openstack/nova: Switch from FLAGS to CONF in nova.console https://review.openstack.org/15361 | 06:51 |
| *** sunilthaha has quit IRC | 06:55 | |
| *** wenjianhn has quit IRC | 06:56 | |
| openstackgerrit | A change was merged to openstack/nova: Switch from FLAGS to CONF in nova.virt https://review.openstack.org/15364 | 06:58 |
| *** samkottler is now known as samkottler|out | 07:00 | |
| *** metral has joined #openstack-dev | 07:06 | |
| *** jcooley has quit IRC | 07:09 | |
| *** eglynn__ has joined #openstack-dev | 07:09 | |
| *** eglynn_ has quit IRC | 07:11 | |
| *** Mandell has joined #openstack-dev | 07:12 | |
| *** mindpixel has joined #openstack-dev | 07:12 | |
| *** metral has quit IRC | 07:16 | |
| openstackgerrit | A change was merged to openstack/cinder: Add admin only action for force detach https://review.openstack.org/14859 | 07:24 |
| *** littleidea has quit IRC | 07:27 | |
| *** markwash has quit IRC | 07:40 | |
| *** Tross has joined #openstack-dev | 07:44 | |
| *** ndipanov has joined #openstack-dev | 07:48 | |
| *** littleidea has joined #openstack-dev | 07:49 | |
| *** AnilV4 has quit IRC | 07:50 | |
| *** AnilV4 has joined #openstack-dev | 07:52 | |
| *** eglynn__ has quit IRC | 07:54 | |
| *** eglynn__ has joined #openstack-dev | 07:55 | |
| *** lloydde has joined #openstack-dev | 07:57 | |
| *** FlaPer87 has joined #openstack-dev | 07:58 | |
| *** lloydde has quit IRC | 08:01 | |
| *** corXi has joined #openstack-dev | 08:03 | |
| *** salv-orlando has joined #openstack-dev | 08:04 | |
| *** EmilienM has joined #openstack-dev | 08:04 | |
| *** Tross has quit IRC | 08:09 | |
| *** darjeeling has quit IRC | 08:09 | |
| *** afazekas has joined #openstack-dev | 08:09 | |
| *** nati_ueno_2 has quit IRC | 08:09 | |
| ttx | lifeless: that would not be arbitrary python, but predefined functions loaded from a trusted PYTHONPATH | 08:10 |
| *** reidrac has joined #openstack-dev | 08:11 | |
| *** jakedahn is now known as jakedahn_zz | 08:11 | |
| *** eglynn__ has quit IRC | 08:12 | |
| *** nati_ueno has joined #openstack-dev | 08:15 | |
| ttx | lifeless: that said, I still have to validate that model, so I certainly welcome your expertise :) | 08:16 |
| lifeless | ttx: so for instance, any predefined function that permits arbitrary code execution (such as callbacks, closures, lambdas, pickles, json (default loader) | 08:17 |
| lifeless | ttx: will be abusable | 08:17 |
| ttx | lifeless: rootwrap would still be the thing you execute as root, let me explain how I planned to do it | 08:18 |
| ttx | nova would call (as the nova user): sudo nova-rootwrap py function1 arg1 arg2 | 08:19 |
| *** littleidea has quit IRC | 08:19 | |
| ttx | sudo escalates that to root | 08:19 |
| ttx | nova-rootwrap has a dictionary mapping function1 to a specific module in its trusted PYTHONPATH | 08:19 |
| ttx | arg1 and arg2 would be passed as function arguments as strings | 08:19 |
| ttx | example at end | 08:20 |
| lifeless | ah | 08:20 |
| ttx | sudo nova-rootwrap py my_chown root /etc/moo | 08:20 |
| ttx | rootwrap maps "my_chown" to oslo.rootwrap.functions.my_$chown | 08:21 |
| *** egallen has joined #openstack-dev | 08:21 | |
| ttx | and call is like oslo.rootwrap.functions.my_chown(str(arg1), str(arg2)) | 08:21 |
| ttx | then it's up to my_chown to check parameters | 08:22 |
| ttx | it's not arbitrary python, it's predefined python functions (rather than predefined shell commands) | 08:22 |
| *** Mandell has quit IRC | 08:22 | |
| lifeless | so I dunno, it seems strictly redundant with a setup.py entrypoint that says 'oslo.rootwrap.functions:my_chown' | 08:22 |
| *** diganta has quit IRC | 08:22 | |
| lifeless | which will install into a stub binary | 08:23 |
| ttx | hmm, an entrypoint in... nova ? | 08:23 |
| lifeless | and its more complex because you're reinventing the marshalling issue (which distribute etc entry points offer) | 08:23 |
| doude | danwent: ping | 08:24 |
| lifeless | ttx: http://reinout.vanrees.org/weblog/2010/01/06/zest-releaser-entry-points.html seems like a reasonable quick primer | 08:24 |
| ttx | you still need to escalate to root at one point, which is what sudo nova-rootwrap does | 08:24 |
| ttx | (or would that entrypoint just be on rootwrap side ?) | 08:25 |
| ttx | thx for pointer on entrypoints, will read | 08:25 |
| ttx | the idea being, rootwrap is the only thing you allow to run as root via sudoers | 08:26 |
| ttx | if the function map within rootwrap can use entrypoints instead of a dictionary, sounds good | 08:27 |
| doude | danwent: Oops, sorry. I wanted to ping dansmith, not you :$ | 08:28 |
| doude | dansmith: ping | 08:28 |
| *** zhuadl has joined #openstack-dev | 08:29 | |
| *** darjeeling has joined #openstack-dev | 08:30 | |
| *** apevec has joined #openstack-dev | 08:30 | |
| *** apevec has joined #openstack-dev | 08:30 | |
| lifeless | ttx: my point is that rootwrap doesn't need to special case python code | 08:31 |
| *** eglynn__ has joined #openstack-dev | 08:31 | |
| lifeless | ttx: you can put the special case in setup.py with a one-liner, and keep rootwrap simpler. | 08:32 |
| ttx | lifeless: interesting. I'll definitely look into that! fwiw, I need to work on oslo-ification and a few other things first, so if it's in grizzly, will be late in grizzly. | 08:32 |
| lifeless | sure | 08:33 |
| lifeless | ttx: for clarity, I'm saying, rootwrap still just calls binaries, but we can make it -super- simple to create those binaries | 08:33 |
| *** pixelbeat has joined #openstack-dev | 08:36 | |
| *** salv-orlando has quit IRC | 08:36 | |
| *** Tross has joined #openstack-dev | 08:37 | |
| *** egallen has quit IRC | 08:37 | |
| jeblair | I'm going to restart jenkins to make a configuration change | 08:38 |
| *** openstackjenkins has quit IRC | 08:39 | |
| *** diganta has joined #openstack-dev | 08:41 | |
| *** openstackjenkins has joined #openstack-dev | 08:42 | |
| *** rburon has joined #openstack-dev | 08:43 | |
| *** Dorogs has joined #openstack-dev | 08:44 | |
| *** Gordonz has quit IRC | 08:45 | |
| *** danwent has quit IRC | 08:45 | |
| *** eglynn__ has quit IRC | 08:45 | |
| *** Gordonz has joined #openstack-dev | 08:46 | |
| jeblair | Jenkins is back up | 08:46 |
| *** crandquist has quit IRC | 08:47 | |
| *** darjeeling has quit IRC | 08:49 | |
| *** rushiagr has quit IRC | 08:55 | |
| *** rushiagr has joined #openstack-dev | 08:56 | |
| *** Tross has quit IRC | 08:57 | |
| *** arosen1 has joined #openstack-dev | 08:58 | |
| arosen1 | hey garyk you there? | 08:58 |
| garyk | arosen: hi | 08:59 |
| arosen1 | I see you just +2'ed this (https://review.openstack.org/#/c/15609/) , I just added a comment i'm not sure we want to merge that yet | 08:59 |
| garyk | arosen1: hi | 08:59 |
| *** reed has quit IRC | 08:59 | |
| garyk | arosen1: give me a sec | 09:00 |
| arosen1 | You can still remove your approved before it merges though. | 09:00 |
| arosen1 | k | 09:00 |
| garyk | arosen1: i'll try | 09:01 |
| garyk | arosen1: good cath. hope it is not too late. sorry\ | 09:01 |
| arosen1 | garyk: unless you think it's okay to merge. Though I don't think changing the perm on those files matter unless we have the #! at the top | 09:01 |
| arosen1 | garyk: no worries. I had +2 but then noticed that. | 09:01 |
| *** egallen has joined #openstack-dev | 09:02 | |
| garyk | arosen1: i must be honest i did not see any reason for the patch as there are the binaries that invoke this code. i just thought that it is harmless. | 09:02 |
| *** doude has quit IRC | 09:03 | |
| *** reed has joined #openstack-dev | 09:03 | |
| arosen1 | garyk: I agree | 09:04 |
| garyk | arosen1: how you keeping? | 09:05 |
| *** doude has joined #openstack-dev | 09:05 | |
| arosen1 | garyk: pretty good. Just been really really busy working on a lot of internal stuff. | 09:07 |
| arosen1 | garyk: what about you? | 09:07 |
| *** eglynn__ has joined #openstack-dev | 09:08 | |
| garyk | arosen1: sadly busy with bugs. trying to answer an email on the list which is taking me forever. in dire need of a vacation | 09:08 |
| arosen1 | garyk: yup, thanksgiving is coming soon :) | 09:09 |
| garyk | arosen1: not on our side of the world :) | 09:09 |
| *** afazekas has quit IRC | 09:09 | |
| *** sagar_nikam has joined #openstack-dev | 09:10 | |
| arosen1 | that's to bad :P | 09:11 |
| *** afazekas has joined #openstack-dev | 09:21 | |
| *** zhuadl has quit IRC | 09:22 | |
| *** ondergetekende has joined #openstack-dev | 09:25 | |
| *** darraghb has joined #openstack-dev | 09:26 | |
| *** reed has quit IRC | 09:27 | |
| *** nati_ueno_2 has joined #openstack-dev | 09:28 | |
| *** nati_ueno has quit IRC | 09:30 | |
| *** marcus has joined #openstack-dev | 09:30 | |
| *** marcus has left #openstack-dev | 09:32 | |
| *** eglynn__ has quit IRC | 09:32 | |
| *** marcus has joined #openstack-dev | 09:33 | |
| *** marcus has quit IRC | 09:33 | |
| *** marcus has joined #openstack-dev | 09:33 | |
| *** xchu has quit IRC | 09:33 | |
| *** Dorogs has quit IRC | 09:34 | |
| *** marcus has left #openstack-dev | 09:34 | |
| *** marcus has joined #openstack-dev | 09:34 | |
| *** Dorogs has joined #openstack-dev | 09:40 | |
| *** armaan has joined #openstack-dev | 09:42 | |
| *** eglynn__ has joined #openstack-dev | 09:42 | |
| *** armaan has left #openstack-dev | 09:42 | |
| *** reed has joined #openstack-dev | 09:43 | |
| marcus | hey everyone, i'm currently trying to setup a nova development environment on my machine. So, I followed the instructions on this website http://docs.openstack.org/developer/nova/devref/development.environment.html#mac-os-x-systems. I am using Mac OS X 10.8.2 but getting some unexpected problems during install process. the tools/install_env.py script stopped whens it's dealing with mysql; fatal error: 'my_config.h' file not fou | 09:43 |
| marcus | After that I installed mysql via mac ports but didn't help. Have someone an idea. thank u | 09:43 |
| FlaPer87 | marcus have you tried to locate that header in your system? Did you installed development files? | 09:46 |
| marcus | there is no my_config.h on my system, which development files do u mean? mysql-dev files? | 09:48 |
| FlaPer87 | AFAIK, macports doesn't install lib's headers unless you explicitly do that (*-devel packages) | 09:48 |
| FlaPer87 | (btw, brew is way cleaner than macports) | 09:48 |
| *** afazekas has quit IRC | 09:49 | |
| *** jpich has joined #openstack-dev | 09:52 | |
| FlaPer87 | marcus ^ | 09:53 |
| marcus | thank you, i'll install the mysql dev thing an than retry to install_env | 09:54 |
| *** maploin has joined #openstack-dev | 09:55 | |
| *** maploin has quit IRC | 09:55 | |
| *** maploin has joined #openstack-dev | 09:55 | |
| *** marcus has quit IRC | 09:57 | |
| *** FlaPer87 has quit IRC | 09:58 | |
| diganta | can anyone help me regarding Nova project , ? | 09:59 |
| *** eglynn__ is now known as eglynn | 09:59 | |
| *** Tross has joined #openstack-dev | 10:01 | |
| *** darjeeling has joined #openstack-dev | 10:03 | |
| *** e1mer has quit IRC | 10:04 | |
| *** \etc\bin has joined #openstack-dev | 10:04 | |
| *** Tross has quit IRC | 10:09 | |
| *** \etc\bin is now known as e1mer | 10:10 | |
| diganta | i just want to know the column of volume table | 10:11 |
| *** gongysh has joined #openstack-dev | 10:11 | |
| *** nati_ueno_2 has quit IRC | 10:12 | |
| *** winston-d has joined #openstack-dev | 10:13 | |
| *** nati_ueno has joined #openstack-dev | 10:13 | |
| *** afazekas has joined #openstack-dev | 10:13 | |
| diganta | zbitter: r u there ? | 10:14 |
| zbitter | uh, yes | 10:15 |
| rushiagr | diganta: are you talking about cinder? | 10:15 |
| *** zbitter is now known as zaneb | 10:16 | |
| diganta | no actually i found a bug in Nova , i want to fix , but i need some info , column details of Volume table , | 10:16 |
| rushiagr | diganta: have a look at class Volume in nova/db/sqlalchemy/models.py | 10:18 |
| rushiagr | i hope this is correct way to look at it. someone plz confirm | 10:18 |
| *** mmagr has joined #openstack-dev | 10:18 | |
| *** arata has quit IRC | 10:24 | |
| diganta | ok rushiagr: i am looking at Volume class as u told | 10:26 |
| *** simonmcc has joined #openstack-dev | 10:30 | |
| *** nati_ueno has quit IRC | 10:30 | |
| *** arata has joined #openstack-dev | 10:31 | |
| *** nati_ueno has joined #openstack-dev | 10:31 | |
| *** mindpixel has quit IRC | 10:32 | |
| *** marcus has joined #openstack-dev | 10:32 | |
| *** maxiz has quit IRC | 10:36 | |
| *** ociuhandu has quit IRC | 10:38 | |
| *** corXi has quit IRC | 10:46 | |
| *** markmc has joined #openstack-dev | 10:48 | |
| *** armaan has joined #openstack-dev | 10:48 | |
| *** armaan has left #openstack-dev | 10:48 | |
| *** maurosr has joined #openstack-dev | 10:48 | |
| doude | Hi all, I worked on bug correction for a few days. I wanted to commit my changes but meanwhile the master files had big modifications from another commit | 10:50 |
| doude | and my code gets a lot of conflicts when it rebase with the master | 10:51 |
| doude | how I can resolve that | 10:51 |
| doude | ? | 10:51 |
| doude | is it possible to abandon all my changes in branch and get the master code ? | 10:52 |
| rushiagr | doude: http://wiki.openstack.org/GerritWorkflow | 10:52 |
| rushiagr | have a look at this quick reference guide | 10:52 |
| *** gongysh has quit IRC | 10:52 | |
| rushiagr | it contains a solution to your problem | 10:53 |
| *** openstack has joined #openstack-dev | 18:39 | |
| harlowja | russellb jaypipes look at line 210 of https://review.openstack.org/#/c/13166/16/nova/virt/images.py | 18:39 |
| smoser | but as i said, i'll fix it. but its pain to SRU it. | 18:39 |
| harlowja | then the same line in nova is still there :-p | 18:39 |
| clarkb | basiclly devstack gate does insecure volumes, but that is probably ok | 18:39 |
| alexpilotti | smoser: c.) teh code is in cloud-init since october, so we don't ghave images for Hyper-V now. Existing ones for other hypervisors require anyway work | 18:40 |
| davidkranz | clarkb: Yeah, that makes sense. Thanks for the update. | 18:40 |
| alexpilotti | smoser: a.) I'd gladly provide the fix | 18:40 |
| jgriffith | davidkranz: we may have quited the dd issue anyway (even without the disable patch) | 18:41 |
| alexpilotti | smoser: b.) that could be an issue | 18:41 |
| davidkranz | jgriffith: Cool. | 18:41 |
| jgriffith | davidkranz: clarkb although I haven't been able to confirm in the devstack gate runs that the disable is actually happening | 18:41 |
| *** eglynn has joined #openstack-dev | 18:41 | |
| * jgriffith I should probably look into that | 18:41 | |
| harlowja | jaypipes russellb nm, that code is on trunk | 18:41 |
| *** gatuus has quit IRC | 18:43 | |
| *** kbringard has quit IRC | 18:43 | |
| harlowja | jaypipes found it! | 18:43 |
| harlowja | i'm so smart | 18:43 |
| harlowja | lol | 18:43 |
| jaypipes | hehe | 18:43 |
| *** Lumiere has quit IRC | 18:43 | |
| alexpilotti | smoser: tx for clarifying, I'm going to work on the VHD side | 18:44 |
| *** kbringard has joined #openstack-dev | 18:44 | |
| *** Lumiere has joined #openstack-dev | 18:44 | |
| *** Lumiere has joined #openstack-dev | 18:44 | |
| smoser | alexpilotti, SRU is not impossible. | 18:45 |
| smoser | its just aPITA | 18:45 |
| alexpilotti | smoser: form what I read in your conversation with ociuhandu | 18:46 |
| alexpilotti | smoser: also Xen would hit the same issue | 18:46 |
| *** thovden has joined #openstack-dev | 18:46 | |
| alexpilotti | smoser: do you think that in that case we could go for the SRU? | 18:47 |
| alexpilotti | smoser: it's not an urgent issue for us, as we are going to distribute a custom package for existing distros anyway | 18:48 |
| harlowja | jaypipes: https://review.openstack.org/15655 | 18:49 |
| harlowja | brb | 18:49 |
| *** arbrandes has quit IRC | 18:49 | |
| *** rushiagr has left #openstack-dev | 18:50 | |
| *** thovden has quit IRC | 18:50 | |
| *** jcru|away is now known as jcru | 18:50 | |
| smoser | alexpilotti, well xen just has the same issue as you i think. i honestly dont know though. | 18:51 |
| smoser | i will end up wanting to SRU config-drive-v2 support to 12.04 at some point also | 18:51 |
| *** cp16net|away is now known as cp16net | 18:51 | |
| smoser | so i cant completely avoid that pain. | 18:51 |
| smoser | but i really do think its better for everyone if hypervisors provide consistent behavior where possible. | 18:52 |
| alexpilotti | smoser: I can for sure guarantee a beer at the next summit :-) | 18:52 |
| smoser | can we just do both? | 18:52 |
| *** m4xmr has joined #openstack-dev | 18:52 | |
| *** m4xmr has quit IRC | 18:53 | |
| *** s0mik has joined #openstack-dev | 18:53 | |
| alexpilotti | smoser: both as the "disk ISO" and "cdrom ISO" approach? | 18:53 |
| *** tmishler has quit IRC | 18:54 | |
| smoser | as in hyper-v would attach disk, but cloud-init would relax its insistance on disk. | 18:54 |
| *** tmishler has joined #openstack-dev | 18:54 | |
| smoser | so that in the future i can swear at myself for relaxing it, when it becomes a problem (i dont know why). | 18:54 |
| *** eglynn has quit IRC | 18:55 | |
| *** m4xmr has joined #openstack-dev | 18:55 | |
| *** colinmcnamara has joined #openstack-dev | 18:55 | |
| *** m4xmr has quit IRC | 18:55 | |
| alexpilotti | smoser: what do you think if we commit the "cdrom" option now, adding the disk option later (e.g. g-2)? | 18:56 |
| *** hemna has quit IRC | 18:57 | |
| *** esp2 has joined #openstack-dev | 18:58 | |
| smoser | sure. as long as "adding disk option later" != "get smoser to stop complaining" | 18:58 |
| *** gatuus has joined #openstack-dev | 18:59 | |
| *** SlickNik has quit IRC | 19:00 | |
| alexpilotti | smoser: sure. We have only an issue to solve for that: dd'ing the iso to a VHD requires mounting the VHD as a loopback device. And that could create race condition issues, as there's no way to identify the mounted device beside iterating on teh volumes | 19:01 |
| *** SlickNik has joined #openstack-dev | 19:01 | |
| *** danwent has quit IRC | 19:01 | |
| smoser | alexpilotti, do you really not have any way to convert a "raw" file to a vhd ? | 19:01 |
| smoser | i'm genuinely asking. | 19:02 |
| smoser | 'qemu-img convert' reports to support vhd | 19:02 |
| alexpilotti | smoser: I'm thinking about qemu-img convert | 19:02 |
| alexpilotti | smoser: but I have to check the licensing | 19:02 |
| smoser | i'm pretty sure it doesnt support all the new-fangled vhd goodness | 19:02 |
| alexpilotti | smoser: it does indeed | 19:02 |
| alexpilotti | smoser: we are using it in other projects to convert VMDKs to VHDs | 19:03 |
| alexpilotti | smoser: the only pity is that it doesn't support AMI but that's another story | 19:03 |
| *** mnewby_ has joined #openstack-dev | 19:03 | |
| *** mnewby_ has quit IRC | 19:03 | |
| *** networkstatic has quit IRC | 19:03 | |
| smoser | what is "ami" ? | 19:04 |
| *** mnewby_ has joined #openstack-dev | 19:04 | |
| alexpilotti | smoser: the AWS format | 19:04 |
| smoser | alexpilotti, also, there is a tool in virtualbox that can do this stuff. | 19:04 |
| smoser | i dont know what na aws format is. | 19:04 |
| *** crandquist has joined #openstack-dev | 19:04 | |
| smoser | its just 'raw' for them. | 19:04 |
| *** spiffxp is now known as spiffxp-test | 19:05 | |
| *** gatuus has quit IRC | 19:05 | |
| alexpilotti | smoser: also we have to port quemu-img to Windows in that case | 19:05 |
| *** mnewby has quit IRC | 19:06 | |
| *** kobier has joined #openstack-dev | 19:06 | |
| *** mnewby_ is now known as mnewby | 19:06 | |
| *** crandquist has quit IRC | 19:06 | |
| *** andrewbogott_afk is now known as andrewbogott | 19:07 | |
| *** maurosr has quit IRC | 19:10 | |
| *** maurosr has joined #openstack-dev | 19:10 | |
| *** metral has quit IRC | 19:10 | |
| alexpilotti | smoser: I checked, quemu is not supporting VHDs | 19:10 |
| *** danwent has joined #openstack-dev | 19:10 | |
| *** networkstatic has joined #openstack-dev | 19:10 | |
| *** annashen has joined #openstack-dev | 19:12 | |
| *** saurabh__ has joined #openstack-dev | 19:13 | |
| *** saurabh__ has left #openstack-dev | 19:13 | |
| smoser | it does support *something* called vhd | 19:14 |
| smoser | "vpc VirtualPC compatible image format (VHD)." | 19:14 |
| smoser | and also, check the virtualbox utility | 19:14 |
| *** spiffxp-test is now known as spiffxp | 19:14 | |
| *** worldofmanish has quit IRC | 19:14 | |
| alexpilotti | smoser: yep, we use it to convert VMDKs to VHDs | 19:15 |
| alexpilotti | smoser: but it doesn't support isos | 19:15 |
| alexpilotti | smoser: I'm going to check the virtualbox tool | 19:16 |
| alexpilotti | smoser: the problem there is that it might be impossible to package it with a separate product from a licensing perspective, I have to check | 19:16 |
| *** hemna has joined #openstack-dev | 19:21 | |
| smoser | alexpilotti, but an iso is just "raw" , no? | 19:22 |
| *** gatuus has joined #openstack-dev | 19:22 | |
| alexpilotti | smoser: good point :-) | 19:23 |
| *** saurabh_ has joined #openstack-dev | 19:23 | |
| jgriffith | anybody from stable-maint-team available to approve https://review.openstack.org/#/c/15594/ | 19:23 |
| alexpilotti | smoser: I'm going to do a test right away | 19:23 |
| notmyname | jaypipes: commented on swift tests (tl;dr: -1 needs work) | 19:24 |
| *** SlickNik has quit IRC | 19:25 | |
| smoser | alexpilotti, i have to go now. i'll check back in later. | 19:25 |
| *** SlickNik has joined #openstack-dev | 19:26 | |
| alexpilotti | smoser: ok tx! | 19:26 |
| smoser | thanks for being open to input. | 19:26 |
| *** melwitt has joined #openstack-dev | 19:26 | |
| smoser | :) | 19:26 |
| jaypipes | notmyname: thx! | 19:28 |
| *** egallen has quit IRC | 19:29 | |
| *** rkukura has quit IRC | 19:32 | |
| *** saurabh_ has quit IRC | 19:32 | |
| *** rgoodwin has joined #openstack-dev | 19:33 | |
| *** ayoung has quit IRC | 19:34 | |
| *** SlickNik has quit IRC | 19:35 | |
| *** colinmcnamara has quit IRC | 19:37 | |
| *** SlickNik has joined #openstack-dev | 19:37 | |
| *** openstack has joined #openstack-dev | 19:45 | |
| jaypipes | davidkranz: link to that successful run? | 19:45 |
| *** arata has quit IRC | 19:45 | |
| davidkranz | jaypipes: Look at https://jenkins.openstack.org/view/Bitrot/job/periodic-tempest-devstack-vm-full/7/console from yesterday. | 19:46 |
| davidkranz | jaypipes: Just the seven errors caused by the keystone bug, nova qemu bug and the one I just fixed. | 19:46 |
| jaypipes | davidkranz: the image client timeouts are from the infinite loop bug, not qemu | 19:47 |
| davidkranz | jaypipes: Oh. Then there are two bugs affecting the same tests :( | 19:47 |
| *** alexpilotti has quit IRC | 19:47 | |
| jaypipes | davidkranz: BTW, the KeyError: 'extra' failures... is that due to a change in Keystone? | 19:47 |
| dolphm | jaypipes: yes | 19:48 |
| davidkranz | jaypipes: Yes. I filed a bug and it was merged earlier today. | 19:48 |
| jaypipes | dolphm: k | 19:48 |
| *** SlickNik has joined #openstack-dev | 19:49 | |
| bcwaldon | davidkranz, jaypipes: I haven't been able to spend time on that bug yet - but you should be able to get around it by setting rabbit_durable_queues to match whatever your queue durability actually is | 19:50 |
| *** echohead has joined #openstack-dev | 19:51 | |
| jaypipes | bcwaldon: is that a new glance-api.conf setting? | 19:52 |
| bcwaldon | nope | 19:52 |
| bcwaldon | this bug has been around for quite a while | 19:52 |
| jaypipes | bcwaldon: no, I mean is that a setting in glance-api.conf -- as opposed to a rabbit setting somewhere? | 19:52 |
| davidkranz | jaypipes: Note also that as recently as Nov 1, before the keystone and nova bugs were introduced, the only failure was the one I just fixed: http://logs.openstack.org/12485/3/check/gate-tempest-devstack-vm/16456/console.html.gz | 19:53 |
| bcwaldon | rabbit_durable_queues is the glance-api.conf setting | 19:53 |
| jaypipes | k, thx | 19:53 |
| jaypipes | davidkranz: interesting... | 19:54 |
| davidkranz | jaypipes: I guess we will see for sure when the nova bug fix is merged and I reverify. | 19:54 |
| *** vipul is now known as vipul|away | 19:54 | |
| *** vipul|away is now known as vipul | 19:54 | |
| jaypipes | k | 19:55 |
| *** novas0x2a|laptop has joined #openstack-dev | 19:57 | |
| *** novas0x2a|laptop has quit IRC | 19:58 | |
| *** novas0x2a|laptop has joined #openstack-dev | 19:58 | |
| *** jcooley has quit IRC | 20:00 | |
| *** alexpilotti has joined #openstack-dev | 20:00 | |
| *** metral has joined #openstack-dev | 20:03 | |
| BLZbubba | is there an easy way to have one of the guest vm's boot up with the qemu CLI option "-std-vga" ? | 20:04 |
| *** maurosr has quit IRC | 20:05 | |
| BLZbubba | i wouldn't mind trying it for all vm's but it may break the existing ones | 20:06 |
| *** danwent has quit IRC | 20:06 | |
| *** danwent has joined #openstack-dev | 20:07 | |
| jgriffith | annegentle_: ping | 20:08 |
| annegentle_ | jgriffith: hey | 20:08 |
| jgriffith | annegentle_: howdy | 20:08 |
| annegentle_ | jgriffith: how goes it? | 20:09 |
| jgriffith | annegentle_: quick question.... | 20:09 |
| annegentle_ | jgriffith: shoot | 20:09 |
| jgriffith | annegentle_: good :) | 20:09 |
| *** krtaylor has quit IRC | 20:09 | |
| jgriffith | annegentle_: I'm now looking at http://docs.openstack.org/trunk/openstack-compute/admin/content/ch_volumes.html | 20:09 |
| *** vipul is now known as vipul|away | 20:09 | |
| annegentle_ | jgriffith: okie | 20:09 |
| jgriffith | annegentle_: appropriate to just linkk "Install cinder" to the other doc? | 20:09 |
| jgriffith | annegentle_: know what I mean there? | 20:10 |
| annegentle_ | jgriffith: might be best, yeah… really we need to figure out if we need a whole new "Volumes with Cinder" type of chapter? | 20:10 |
| jgriffith | annegentle_: I think we do (but not today) :) | 20:10 |
| jgriffith | annegentle_: that brings me to the next questionn | 20:10 |
| annegentle_ | jgriffith: yeah is it a new chapter ? Yeah I think so. | 20:10 |
| jgriffith | annegentle_: yeah, I think it's a new chapter for sure | 20:11 |
| jgriffith | annegentle_: the other question.... | 20:11 |
| annegentle_ | jgriffith: ok | 20:11 |
| annegentle_ | hey I should have asked, is the Storwize stuff for Folsom or only Grizzly? | 20:11 |
| jgriffith | annegentle_: Do we leave the nova-volume ref's in trunk for now or do I start switching those over? | 20:11 |
| * annegentle_ needs to release folsom docs | 20:11 | |
| jgriffith | annegentle_: Folsom | 20:11 |
| *** datsun180b has joined #openstack-dev | 20:11 | |
| annegentle_ | jgriffith: ok, whew. | 20:12 |
| jgriffith | :) | 20:12 |
| annegentle_ | jgriffith: so I need to cut a branch | 20:12 |
| *** networkstatic has quit IRC | 20:12 | |
| annegentle_ | jgriffith: working on it this week | 20:12 |
| annegentle_ | jgriffith: so, leave those in trunk for now and start a new chapter | 20:12 |
| jgriffith | annegentle_: alright coolio | 20:12 |
| jgriffith | annegentle_: thanks! | 20:12 |
| annegentle_ | then in grizzly's new branch (the new trunk I'll make) we'll have a place to really dig in | 20:12 |
| jgriffith | annegentle_: Yup, sounds good | 20:12 |
| jgriffith | annegentle_: I'll hopefully do some doc work this afternoon, then work on some patches tonight | 20:13 |
| annegentle_ | jgriffith: nice | 20:13 |
| jgriffith | annegentle_: Then I'm gone for a few days moving cattle :) | 20:13 |
| jgriffith | annegentle_: Thanks for the help | 20:14 |
| *** roge has quit IRC | 20:14 | |
| *** rdo has quit IRC | 20:14 | |
| annegentle_ | jgriffith: really moving cattle? | 20:15 |
| jgriffith | annegentle_: Yup! :) One of my favorite things | 20:16 |
| jgriffith | annegentle_: once in the spring and once in the fall | 20:16 |
| jgriffith | annegentle_: couple days horseback, no phone, computer etc | 20:16 |
| *** adjohn has quit IRC | 20:16 | |
| davidkranz | annegentle_: api.openstack.org has two broken links. API Complete reference points to self, and API Specifications is as broken link | 20:18 |
| *** jcru is now known as jcru|away | 20:19 | |
| annegentle_ | davidkranz: thanks - moving around some of the Jenkins jobs and source now | 20:19 |
| annegentle_ | davidkranz: will fix asap! | 20:19 |
| BLZbubba | ok so there is no way to change the qemu-kvm command line options for vga? | 20:21 |
| *** roge has joined #openstack-dev | 20:21 | |
| *** alop has joined #openstack-dev | 20:21 | |
| *** krtaylor has joined #openstack-dev | 20:21 | |
| *** aeperezt has quit IRC | 20:22 | |
| *** eglynn has joined #openstack-dev | 20:23 | |
| *** aeperezt has joined #openstack-dev | 20:26 | |
| *** roge has quit IRC | 20:28 | |
| *** jcru|away is now known as jcru | 20:29 | |
| *** Dorogs has joined #openstack-dev | 20:29 | |
| *** winston-d has quit IRC | 20:30 | |
| *** dhellmann has joined #openstack-dev | 20:33 | |
| *** FlaPer87 has quit IRC | 20:34 | |
| *** jcru is now known as jcru|away | 20:34 | |
| *** jcru|away is now known as jcru | 20:37 | |
| *** pixelbeat has quit IRC | 20:37 | |
| *** otherwiseguy has quit IRC | 20:44 | |
| *** armaan has left #openstack-dev | 20:45 | |
| jaypipes | bcwaldon: how do I determine what the durability of my rabbit queues are if devstack if setting it all up? | 20:46 |
| *** metral has quit IRC | 20:47 | |
| *** arata has joined #openstack-dev | 20:47 | |
| lifeless | jaypipes: rabbitmq-ctl list_queues | 20:47 |
| lifeless | jaypipes: rabbitmq-ctl list_exchanges | 20:47 |
| *** otherwiseguy has joined #openstack-dev | 20:47 | |
| *** hemna has quit IRC | 20:48 | |
| jaypipes | lifeless: how does that help determine the durability? | 20:48 |
| lifeless | it should tell you | 20:48 |
| lifeless | its a flag on the queue / exchange | 20:48 |
| lifeless | might need the admin cli; one sec | 20:49 |
| *** danpb has joined #openstack-dev | 20:49 | |
| *** metral_ has joined #openstack-dev | 20:49 | |
| jaypipes | lifeless: am I missing something here? http://paste.openstack.org/show/25089/ | 20:49 |
| *** Tross has joined #openstack-dev | 20:50 | |
| lifeless | list_queues name durable auto_delete messages consumers | 20:51 |
| lifeless | list_exchanges name type durable auto_delete arguments | 20:52 |
| lifeless | jaypipes: ^ | 20:52 |
| *** egallen has joined #openstack-dev | 20:52 | |
| *** adjohn has joined #openstack-dev | 20:54 | |
| openstackgerrit | A change was merged to openstack/horizon: Flavor Extra Specs support. https://review.openstack.org/13169 | 20:54 |
| bcwaldon | jaypipes: well, this isn't an intelligent way to do it, but if you see that bug then invert the value of rabbit_durable_queues ;) | 20:54 |
| jaypipes | bcwaldon: yeah... just trying to figure out how to get devstack to do it... | 20:55 |
| jaypipes | lifeless: much appreciated... got it. | 20:55 |
| jaypipes | http://paste.openstack.org/show/25090/ | 20:55 |
| jaypipes | so there's my problem :) just need to get devstack to write out True, not False.. | 20:55 |
| *** Ruetobas has quit IRC | 20:57 | |
| openstackgerrit | A change was merged to openstack/glance: Ensure authorization before deleting from store https://review.openstack.org/15658 | 20:57 |
| *** egallen has quit IRC | 20:58 | |
| *** egallen has joined #openstack-dev | 20:58 | |
| mnewby | danwent: *ping* | 20:58 |
| danwent | mnewby: hey | 20:59 |
| lifeless | danwent: hi, you are heavily involved in quantum, right ? | 20:59 |
| mnewby | danwent: Hi Dan. | 20:59 |
| danwent | mnewby: hey | 20:59 |
| danwent | lifeless: yes, one could say that :) | 20:59 |
| vishy | nova meeting about to start in #openstack-meeting | 20:59 |
| mnewby | danwent: I'm looking at the quantum packaging for ubuntu and for some reason the ovs plugin package depends on openvswitch-switch. Shouldn't that only be required for the agent? | 20:59 |
| lifeless | danwent: If so, I have a somewhat random question for you - is anyone successfully running quantum w/openflow on top of quantum w/openflow. | 21:00 |
| lifeless | danwent: [also hi, pleased to meet you :)] | 21:00 |
| danwent | mnewby: yes, that would be more accurate | 21:00 |
| mnewby | danwent: Ok, I'll follow up. | 21:00 |
| mnewby | danwent: danke | 21:00 |
| danwent | mnewby: thx | 21:00 |
| danwent | lifeless: depends what you mean by "quantum on openflow". | 21:00 |
| lifeless | danwent: I'm certain I have my terminology wrong. What I'm considering is this: | 21:01 |
| danwent | for example, the nicira solution uses openflow as part of the solution, but there are other solutions that use openflow in a very different way. | 21:01 |
| *** metral has joined #openstack-dev | 21:01 | |
| lifeless | nova baremetal to deploy physical machines. Those machines running on a SDN managed by quantum, implemented by openflow hardware. | 21:01 |
| *** Ruetobas has joined #openstack-dev | 21:02 | |
| *** hemna has joined #openstack-dev | 21:02 | |
| lifeless | then, those machines run nova compute, hosting VM's, and the VM networks are managed by quantum, with ovs on the host and tunnelling | 21:02 |
| lifeless | between compute hosts | 21:02 |
| lifeless | My understanding of the theory is that this should all Just Work, but I'd love to know there is an existence proof out there :) | 21:02 |
| *** jog0_ has joined #openstack-dev | 21:03 | |
| danwent | lifeless: that is certainly conceptually possible. however, the set of openflow enabled hardware is pretty limited right now, so I'm not aware of anyone with a production deployment doing that (not saying there aren't any, I am just not familiar with any of them). | 21:03 |
| danwent | most people just run standard L3 router protocols, and tunnel over them, in my experience. | 21:04 |
| lifeless | ok, thanks. | 21:04 |
| *** nunosantos has joined #openstack-dev | 21:04 | |
| *** metral_ has quit IRC | 21:04 | |
| openstackgerrit | A change was merged to openstack/nova: Limit formatting routes when adding resources https://review.openstack.org/15623 | 21:05 |
| zykes- | danwent: openflow in production on hardware except the big G | 21:05 |
| *** rburon has quit IRC | 21:05 | |
| *** jog0 has quit IRC | 21:05 | |
| *** jog0_ is now known as jog0 | 21:05 | |
| *** anniec has quit IRC | 21:06 | |
| danwent | zykes-: hehe, yes, i was limiting myself to discussing openstack deployments. What G did would be well beyond what most other companies could do. | 21:06 |
| zykes- | :p | 21:06 |
| openstackgerrit | A change was merged to openstack/horizon: Wrap a project list call with exception handler. https://review.openstack.org/15439 | 21:07 |
| *** egallen has quit IRC | 21:07 | |
| *** egallen has joined #openstack-dev | 21:07 | |
| *** egallen has quit IRC | 21:08 | |
| *** anniec has joined #openstack-dev | 21:10 | |
| *** metral has quit IRC | 21:10 | |
| *** epim has quit IRC | 21:15 | |
| *** dspano_ has joined #openstack-dev | 21:17 | |
| *** Dorogs has quit IRC | 21:18 | |
| *** boden has quit IRC | 21:19 | |
| *** s0mik has quit IRC | 21:19 | |
| *** littleidea has joined #openstack-dev | 21:19 | |
| *** s0mik has joined #openstack-dev | 21:22 | |
| *** radez is now known as radez_g0n3 | 21:24 | |
| *** gyee has joined #openstack-dev | 21:29 | |
| *** s0mik has quit IRC | 21:30 | |
| *** cp16net is now known as cp16net|away | 21:32 | |
| *** afazekas has quit IRC | 21:34 | |
| bcwaldon | jeblair: ping | 21:36 |
| *** vipul|away is now known as vipul | 21:37 | |
| zykes- | mikal: around ? | 21:37 |
| *** dolphm has quit IRC | 21:37 | |
| *** cp16net|away is now known as cp16net | 21:38 | |
| *** roge has joined #openstack-dev | 21:39 | |
| *** danpb has quit IRC | 21:44 | |
| *** ayoung has joined #openstack-dev | 21:47 | |
| *** juice has joined #openstack-dev | 21:47 | |
| *** corXi has quit IRC | 21:48 | |
| *** littleidea has quit IRC | 21:48 | |
| *** rnirmal has quit IRC | 21:51 | |
| *** epopt37 has joined #openstack-dev | 21:53 | |
| *** littleidea has joined #openstack-dev | 21:57 | |
| *** Dorogs has joined #openstack-dev | 21:58 | |
| *** Dorogs has quit IRC | 21:59 | |
| *** littleidea has quit IRC | 22:00 | |
| *** markmcclain1 has joined #openstack-dev | 22:01 | |
| *** jcooley has joined #openstack-dev | 22:01 | |
| *** markvoelker has quit IRC | 22:01 | |
| *** markmcclain has quit IRC | 22:02 | |
| *** jog0 has quit IRC | 22:02 | |
| *** jog0 has joined #openstack-dev | 22:03 | |
| *** littleidea has joined #openstack-dev | 22:03 | |
| *** anniec has quit IRC | 22:03 | |
| davidkranz | sdague, dansmith : A new test was added to this https://review.openstack.org/#/c/15655/2 | 22:04 |
| *** anniec has joined #openstack-dev | 22:04 | |
| clarkb | bcwaldon: jeblair is in Europe so you probably won't hear from him for many hours. Anything the CI team can help with? | 22:05 |
| dansmith | davidkranz: looking | 22:05 |
| dansmith | davidkranz: done | 22:06 |
| davidkranz | dansmith: Thanks. Now we will see if this fixes the tempest problem. | 22:07 |
| bcwaldon | clarkb: yes, I never know which one of you guys to poke | 22:07 |
| bcwaldon | clarkb: this might need you to push a button on the backend - https://review.openstack.org/#/c/15142/ | 22:07 |
| bcwaldon | clarkb: its cleared to land but it hasnt been merged | 22:07 |
| clarkb | bcwaldon: apparently it needs a recursive merge to get merged which gerrit won't do | 22:08 |
| bcwaldon | clarkb: should I just rebase it and re-approve? | 22:08 |
| clarkb | bcwaldon: yes, that should get it through | 22:08 |
| bcwaldon | k, thanks! | 22:08 |
| clarkb | if not poke me again and we'll dig into it more | 22:09 |
| *** mtreinish has quit IRC | 22:10 | |
| *** dolphm has joined #openstack-dev | 22:11 | |
| openstackgerrit | A change was merged to openstack/python-novaclient: show help when calling without arguments https://review.openstack.org/15665 | 22:11 |
| *** kbringard has quit IRC | 22:11 | |
| *** koolhead17 has joined #openstack-dev | 22:13 | |
| *** saurabhsurana_ has joined #openstack-dev | 22:13 | |
| *** PaulM has quit IRC | 22:14 | |
| *** PaulM has joined #openstack-dev | 22:16 | |
| *** e1mer has joined #openstack-dev | 22:17 | |
| vishy | russellb: that eventlet issue is interesting | 22:18 |
| russellb | vishy: it made me want to rage | 22:18 |
| russellb | i seriously saw select() saying a pipe was ready for reading when nothing had written to it | 22:18 |
| russellb | had lots of logging to verify it | 22:19 |
| russellb | but i can't reproduce it with simple test code | 22:19 |
| *** tgall_foo has quit IRC | 22:25 | |
| vishy | russellb: gross | 22:25 |
| vishy | russellb: I ran into a really annoying eventlet issue the other day | 22:25 |
| vishy | where the return code from popen was incorrect | 22:25 |
| russellb | awesome.. | 22:26 |
| russellb | did you see my patch to wrap eventlet's select? heh | 22:26 |
| vishy | russellb: still didn't figure out how it was happening | 22:26 |
| russellb | i got the result from its select() then did a non-blocking call to the built-in select to sanity check the result | 22:26 |
| russellb | (i abandoned it, felt way too dirty about it) | 22:26 |
| vishy | russellb: but apparently not monkeypatching os does it. | 22:26 |
| russellb | yeah, so you're not going through eventlet's version | 22:27 |
| russellb | same with this bug, not patching select fixes it | 22:27 |
| russellb | that's what we get for replacing core stdlib stuff, heh | 22:27 |
| lifeless | random question time | 22:28 |
| lifeless | why do we use eventlet ? | 22:28 |
| russellb | watermelons! | 22:28 |
| russellb | (random answer time) | 22:28 |
| lifeless | it seems super fragile | 22:28 |
| vishy | lifeless: alternative solution? | 22:29 |
| lifeless | vishy: to what problem definition ? | 22:29 |
| vishy | lifeless: async code | 22:29 |
| lifeless | vishy: I'm not criticising the choice, I'm asking what problems we're solving w/eventlet I guess. | 22:29 |
| russellb | so what's wrong with regular threads? | 22:29 |
| *** kgriffs has quit IRC | 22:29 | |
| vishy | russellb: locking is hard :) | 22:30 |
| russellb | ha | 22:30 |
| vishy | russellb: actually it is just another set of problems | 22:30 |
| PaulM | so we used eventlet. hah. | 22:30 |
| *** metral has joined #openstack-dev | 22:30 | |
| PaulM | now we have 2 problems | 22:30 |
| vishy | PaulM: actually we started with twisted | 22:30 |
| PaulM | oh | 22:31 |
| * russellb would rather have real threads and do locking | 22:31 | |
| PaulM | well, that might be an upgrade | 22:31 |
| *** sniperd has quit IRC | 22:31 | |
| vishy | and everyone was like, but eventlet is so much easier! | 22:31 |
| russellb | we're not doing *that* much data sharing | 22:31 |
| lifeless | vishy: so, were they right ? | 22:31 |
| vishy | easier to use, yes, by far | 22:31 |
| vishy | less buggy? questionable? | 22:31 |
| PaulM | russellb: I agree. Eventlet is unnecessary | 22:31 |
| lifeless | vishy: :) 'better overall' I think is the metric I'm interested in. | 22:32 |
| russellb | but who has "work on eventlet alternative" at the top of their todo list ... | 22:32 |
| lifeless | vishy: whats your take from that angle ? | 22:32 |
| vishy | I actually just think threads / eventlet / multiprocessing / twisted each comes with their own unique difficulties | 22:32 |
| vishy | concurrency is hard in general | 22:32 |
| lifeless | multiprocessing is terribly slow | 22:32 |
| russellb | +1 to that | 22:32 |
| vishy | everyone just complains about the concurrency model they are currently using | 22:32 |
| russellb | to vishy's comment i mean. | 22:32 |
| *** jcooley has quit IRC | 22:32 | |
| vishy | grass is always greener | 22:32 |
| PaulM | sure, but how much data are we sharing? We're running a freakin' rest API, this should not be stateful inside our process | 22:32 |
| russellb | i wasn't nearly as angry until i hit this bug this week :) | 22:32 |
| vishy | the real solution is erlang obv. | 22:32 |
| PaulM | for srs | 22:33 |
| lifeless | vishy: I'm not sure about that. There are lots of folk in various communities that are very happy to be there. | 22:33 |
| lifeless | [that == everyone complains] | 22:33 |
| * russellb came from a heavily multi-threaded app before this, and definitely complained about it :-p | 22:33 | |
| vishy | lifeless: show me a project taht is using the number of different technologies we are using across multiple machines | 22:33 |
| vishy | and is happy with there concurrency model | 22:33 |
| vishy | :) | 22:33 |
| vishy | * their | 22:33 |
| russellb | I DON'T HEAR YOU | 22:34 |
| russellb | stop trying to be so reasonable | 22:34 |
| lifeless | vishy: Launchpad is similar in size and pretty happy, using a combination of tools - threads for application servers, twisted for custom network protocols / backend services. | 22:34 |
| PaulM | vishy: serious question: which parts of the stack are actually using features of eventlet that can't be trivially replicated by separate threaded workers? | 22:35 |
| lifeless | vishy: not to say that that would work elsewhere, just answering the qestion. | 22:35 |
| vishy | PaulM: unknown, but I know there are assumptions in various places that we aren't using real threads | 22:36 |
| russellb | i got flamed for pointing out one of those in a patch recently | 22:36 |
| russellb | and the author told me that it wasn't worth fixing | 22:36 |
| russellb | :-( | 22:36 |
| *** dhellmann has quit IRC | 22:36 | |
| PaulM | vishy: interesting. so we're using eventlet because nobody knows why we can't | 22:37 |
| *** sunilthaha has joined #openstack-dev | 22:37 | |
| openstackgerrit | A change was merged to openstack/horizon: Remove superfluous slash from admin usage.csv template https://review.openstack.org/15142 | 22:37 |
| *** littleidea has quit IRC | 22:38 | |
| PaulM | <sigh> ah well. At least most of the code I touch doesn't have to deal with eventlet | 22:38 |
| russellb | well i don't know that it's that nobody knows | 22:39 |
| russellb | it's working well enough that it hasn't become a top priority for anyone to investigate changing | 22:39 |
| nati_uen_ | +1 for erlang | 22:39 |
| *** rbasak has quit IRC | 22:39 | |
| vishy | PaulM: no we're using eventlet because writing things in threaded code is hard so no one has taken the time to try to rip it out :) | 22:39 |
| *** littleidea has joined #openstack-dev | 22:40 | |
| * russellb likes the challenge! | 22:40 | |
| PaulM | what things? I'm trying to understand an example where we'd be talking between threads | 22:40 |
| vishy | PaulM: although it has been suggested multiple times for the compute worker | 22:40 |
| vishy | PaulM: shared state in the scheduler | 22:40 |
| vishy | PaulM: shared state in the compute worker | 22:41 |
| *** annegentle_ has quit IRC | 22:41 | |
| PaulM | ah, so we did actually write that in | 22:41 |
| vishy | PaulM: the api worker would be fine but on the other hand evented code is pretty good for api code | 22:41 |
| * PaulM is in a war on stateful things | 22:41 | |
| russellb | it's actually more the compute side that we have eventlet trouble | 22:41 |
| *** epim has joined #openstack-dev | 22:42 | |
| vishy | every summit someone suggests turning it into a simple worker | 22:42 |
| vishy | but I still haven't seen code :) | 22:42 |
| *** epim has quit IRC | 22:42 | |
| russellb | heh | 22:42 |
| *** epim has joined #openstack-dev | 22:42 | |
| russellb | dansmith and I have been talking about a nova-compute-proxy that looks more like nova-sink at first, but slowly stuff starts moving up | 22:42 |
| *** Tross has quit IRC | 22:42 | |
| russellb | i don't think we'll get to this really nice think nova-compute some folks envision | 22:43 |
| russellb | but we can set the track that direction | 22:43 |
| russellb | s/think/thin/ | 22:44 |
| russellb | but we'll have real code :) | 22:45 |
| *** markmcclain1 has quit IRC | 22:47 | |
| *** koolhead17 has quit IRC | 22:49 | |
| vishy | yay for real code | 22:50 |
| *** cp16net is now known as cp16net|away | 22:53 | |
| *** cp16net|away is now known as cp16net | 22:53 | |
| *** thovden has joined #openstack-dev | 22:54 | |
| *** echohead has quit IRC | 22:54 | |
| *** thovden has left #openstack-dev | 22:55 | |
| *** echohead has joined #openstack-dev | 22:57 | |
| *** gatuus has quit IRC | 22:58 | |
| *** sunilthaha has quit IRC | 22:58 | |
| *** anniec has quit IRC | 23:01 | |
| *** anniec has joined #openstack-dev | 23:02 | |
| *** vipul is now known as vipul|away | 23:03 | |
| *** FlaPer87 has joined #openstack-dev | 23:04 | |
| *** arosen has left #openstack-dev | 23:08 | |
| *** dspano_ has quit IRC | 23:10 | |
| *** alunduil has quit IRC | 23:10 | |
| russellb | vishy: so, i'm concerned about this patch ... https://review.openstack.org/#/c/15594/ | 23:12 |
| vishy | russellb: that is the one i was talking about! :) | 23:13 |
| russellb | yeah :) | 23:13 |
| russellb | but when i looked it, i realized that preventing eventlet from switching on read() and write() is a little worrying | 23:13 |
| russellb | since those are going to get used a lot | 23:13 |
| russellb | and all the sudden we've removed it | 23:13 |
| russellb | based on the commit message ... is this as simple as doing a file presence check before calling openssl? | 23:14 |
| russellb | guess i need to pull up keystone | 23:14 |
| *** aeperezt has quit IRC | 23:14 | |
| *** gatuus has joined #openstack-dev | 23:15 | |
| *** echohead has quit IRC | 23:15 | |
| jaypipes | davidkranz: yeah, so I fixed the rabbit durable shit and yeah, still getting those errors (timeout) for any test using services/compute/json/images_client.py :( | 23:16 |
| *** hattwick has quit IRC | 23:16 | |
| vishy | russellb: we could change it to do that but i'm worried about the underlying issue | 23:17 |
| russellb | i certainly understand that | 23:17 |
| russellb | tempting to not patch select() either for the same reason, heh' | 23:17 |
| russellb | but ... we're going to kill our threads | 23:18 |
| russellb | i don't have a good answer i guess | 23:18 |
| russellb | can fix this one place where the error occurred, but for the general case, we're in trouble either way | 23:18 |
| russellb | but hurting thread scheduling, or dealing with knowing that eventlet might lie to us | 23:18 |
| russellb | sigh. | 23:18 |
| *** nati_uen_ has quit IRC | 23:19 | |
| *** hattwick has joined #openstack-dev | 23:19 | |
| *** nati_uen_ has joined #openstack-dev | 23:19 | |
| *** jergerber has quit IRC | 23:19 | |
| *** jergerber has joined #openstack-dev | 23:21 | |
| *** aeperezt has joined #openstack-dev | 23:21 | |
| *** sacharya has quit IRC | 23:21 | |
| *** markwash has quit IRC | 23:22 | |
| *** datsun180b has quit IRC | 23:23 | |
| *** imsplitbit has quit IRC | 23:23 | |
| *** matiu has quit IRC | 23:23 | |
| *** zaitcev has quit IRC | 23:23 | |
| *** Hien has quit IRC | 23:23 | |
| *** fungi has quit IRC | 23:23 | |
| *** pixelbeat has joined #openstack-dev | 23:24 | |
| vishy | russellb: I have repro case! | 23:24 |
| *** Hien has joined #openstack-dev | 23:24 | |
| russellb | \o/ | 23:24 |
| *** fungi has joined #openstack-dev | 23:26 | |
| *** zaitcev has joined #openstack-dev | 23:27 | |
| *** gatuus has quit IRC | 23:27 | |
| *** Gordonz has quit IRC | 23:29 | |
| *** jaypipes has quit IRC | 23:30 | |
| vishy | russellb: http://paste.openstack.org/show/25123/ | 23:30 |
| vishy | i think it has to do with openssl closing file handles | 23:30 |
| vishy | trying to repro with a different binary | 23:30 |
| *** jimfehlig has quit IRC | 23:31 | |
| russellb | so here's a question ... why does the code open pipes for stdin and stdout, it doesn't make use of it | 23:32 |
| russellb | does read stderr i guess, so it can log it if it fails | 23:32 |
| *** maurosr has joined #openstack-dev | 23:32 | |
| *** roadTripper has quit IRC | 23:33 | |
| *** sniperd has joined #openstack-dev | 23:33 | |
| dolphm | ayoung: ^ | 23:33 |
| *** stevebake has quit IRC | 23:33 | |
| *** sstent has quit IRC | 23:34 | |
| *** sstent has joined #openstack-dev | 23:34 | |
| vishy | russellb: it gets output and err with process.communicate | 23:34 |
| russellb | doesn't do anything with output | 23:35 |
| vishy | russellb: i just tested with cat and cat foobar and it works fine | 23:35 |
| *** jcru is now known as jcru|away | 23:35 | |
| vishy | russellb: that is true | 23:35 |
| russellb | anyway, that's unrelated ... | 23:35 |
| russellb | nice, i reproduced it with your test code | 23:35 |
| russellb | first few times anyway ... now it's working every time | 23:36 |
| russellb | still happens occasionally ... | 23:36 |
| *** markwash has joined #openstack-dev | 23:38 | |
| *** sniperd has quit IRC | 23:38 | |
| *** cp16net is now known as cp16net|away | 23:39 | |
| russellb | vishy: so check this out ... when that happens, there is a zombie openssl process left around | 23:43 |
| russellb | otherwiseguy: ^ | 23:43 |
| vishy | russellb: ah interesting | 23:43 |
| russellb | i verified it putting a sleep at the end of your test code | 23:43 |
| vishy | russellb: if you update the loop to do like 100 | 23:43 |
| *** nati_ueno has quit IRC | 23:43 | |
| vishy | it will start failing | 23:43 |
| russellb | yeah i did | 23:43 |
| vishy | sleep does what? | 23:44 |
| *** jergerber has quit IRC | 23:44 | |
| russellb | well nothing ... just so i can look at 'ps' output and see the zombie processes | 23:44 |
| *** nati_ueno has joined #openstack-dev | 23:44 | |
| *** esheffield has quit IRC | 23:44 | |
| vishy | strace shows a rt_sigaction(SIGPIPE, {SIG_IGN, [PIPE], SA_RESTORER|SA_RESTART, 0x7f46011984a0}, {SIG_IGN, [PIPE], SA_RESTORER|SA_RESTART, 0x7f46011984a0}, 8) = 0 | 23:44 |
| *** nati_ueno has quit IRC | 23:44 | |
| *** nati_ueno has joined #openstack-dev | 23:45 | |
| *** samkottler is now known as samkottler|food | 23:46 | |
| lifeless | arata: around? Would like to chat about the initrd builder a little | 23:48 |
| arata | lifeless: hello | 23:49 |
| lifeless | arata: hi! | 23:49 |
| vishy | russellb: interesting if you set trace in the last loop and hit c each time | 23:49 |
| vishy | it works | 23:49 |
| openstackgerrit | A change was merged to openstack/swift: charset for default responses https://review.openstack.org/15686 | 23:49 |
| vishy | but if you put a sleep in there it still fails | 23:49 |
| lifeless | arata: so, two things I'd like to chat about; shared github project - like, is it possible for me and the folk I'm working with to [eventually] get commit and review rights to the initrd builder tree where it is, or should we have a shared organisation that both ntt & hp and anyone else that collaborates on this can become members of ? | 23:50 |
| lifeless | arata: [one answer to this might be: "you can become a member of the nttdocomo-openstack github organisation'] | 23:51 |
| *** networkstatic has joined #openstack-dev | 23:51 | |
| lifeless | arata: I assume long term it would become part of openstack with openstack-CI etc, but in the medium term having something we can scale out the development of - and do more granular review and evolving of, would be good. | 23:52 |
| russellb | vishy: been at my desk too long ... need to step away, at least for a bit. :-) good luck | 23:52 |
| arata | lifeless: we can register you as a collabolater | 23:52 |
| vishy | russellb: much simplified repro | 23:52 |
| arata | so that you can push freely | 23:53 |
| *** dolphm has quit IRC | 23:56 | |
| lifeless | arata: ok cool. Whenever you feel we're ready for that, that would be great. | 23:56 |
| *** sniperd has joined #openstack-dev | 23:56 | |
| lifeless | arata: the second question is this - we have a bunch of code that is in the same basic space: making images - disk images of openstack, ready to run. I'd like to consolidate all of this stuff. | 23:57 |
| otherwiseguy | vishy, link to the new repro? | 23:57 |
| lifeless | arata: my proposal is that we: | 23:57 |
| lifeless | - put all the image building logic in the initrd-builder tree | 23:57 |
| *** tmishler has quit IRC | 23:57 | |
| lifeless | - this would be initrd, *and* operating-system image code | 23:58 |
| lifeless | - we include in this tree baseline stuff, but may in future want to have a tree that is just the executable code, and a separate tree that is just configurations. | 23:58 |
| *** rpedde is now known as rpedde_away | 23:58 | |
| vishy | otherwiseguy: http://paste.openstack.org/show/25134/ | 23:59 |
| lifeless | arata: if that sounds good to you, I'll start moving stuff around and send you pull requests. | 23:59 |
| otherwiseguy | vishy: thanks | 23:59 |
| lifeless | arata: I'd like to have something that is packagable by distributions, or at least someting folk can install in /usr/local/bin/* | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!