| *** rkukura has joined #openstack-dev | 00:04 | |
| *** sleepsonzzz is now known as sleepsonthefloor | 00:04 | |
| *** _adjohn has joined #openstack-dev | 00:12 | |
| *** adjohn has quit IRC | 00:15 | |
| *** _adjohn is now known as adjohn | 00:15 | |
| *** andrewsben is now known as andrewsben_zz | 00:16 | |
| *** Ryan_Lane has quit IRC | 00:16 | |
| *** Ryan_Lane1 has joined #openstack-dev | 00:17 | |
| *** sleepsonthefloor is now known as sleepsonzzz | 00:17 | |
| *** issackel_ has quit IRC | 00:19 | |
| *** Ryan_Lane1 is now known as Ryan_Lane | 00:19 | |
| *** Ryan_Lane has joined #openstack-dev | 00:19 | |
| *** sandywalsh has quit IRC | 00:32 | |
| *** asdfasdf has quit IRC | 00:35 | |
| *** nati has joined #openstack-dev | 00:35 | |
| *** kindaopsdevy has quit IRC | 00:36 | |
| *** spiffxp has quit IRC | 00:36 | |
| *** jakedahn is now known as jakedahn_zz | 00:37 | |
| *** crobinso has quit IRC | 00:39 | |
| *** mnewby has joined #openstack-dev | 00:42 | |
| *** sandywalsh has joined #openstack-dev | 00:46 | |
| *** EntropyWorks has joined #openstack-dev | 00:51 | |
| *** EntropyWorks has quit IRC | 00:54 | |
| *** reed has quit IRC | 00:55 | |
| *** jdurgin has quit IRC | 00:58 | |
| *** sandywalsh has quit IRC | 00:58 | |
| *** mdomsch has joined #openstack-dev | 01:04 | |
| *** adjohn has quit IRC | 01:04 | |
| *** maplebed has quit IRC | 01:04 | |
| *** sandywalsh has joined #openstack-dev | 01:14 | |
| *** vincentricci has quit IRC | 01:15 | |
| *** mattray has joined #openstack-dev | 01:19 | |
| *** heckj has quit IRC | 01:33 | |
| *** roge has joined #openstack-dev | 01:35 | |
| *** fsouza has quit IRC | 01:40 | |
| *** novas0x2a|laptop has quit IRC | 01:44 | |
| *** Mandell has quit IRC | 01:44 | |
| *** devananda has quit IRC | 01:51 | |
| *** dtroyer_zzz is now known as dtroyer | 02:01 | |
| *** Ryan_Lane has quit IRC | 02:11 | |
| *** davlaps has quit IRC | 02:12 | |
| *** thickskin has joined #openstack-dev | 02:14 | |
| *** fsouza has joined #openstack-dev | 02:19 | |
| *** winston-d has joined #openstack-dev | 02:19 | |
| winston-d | jaypipes, around? | 02:19 |
|---|---|---|
| *** adjohn has joined #openstack-dev | 02:24 | |
| *** bsas has joined #openstack-dev | 02:27 | |
| *** ayoung has quit IRC | 02:27 | |
| bsas | heya, I am trying to build a devstack box | 02:28 |
| bsas | So I installed Ubuntu 11.10 and executed the http://devstack.org/ script | 02:28 |
| bsas | and everything was working but after I rebooted the machine Keystone is not working anymore. | 02:28 |
| bsas | Anyone had that issue? | 02:28 |
| bsas | I just did the basic install, no change... | 02:28 |
| *** thickskin has left #openstack-dev | 02:28 | |
| *** mjfork has quit IRC | 02:31 | |
| *** thickskin has joined #openstack-dev | 02:32 | |
| *** devananda has joined #openstack-dev | 02:32 | |
| *** bsas has left #openstack-dev | 02:39 | |
| *** slyphon has joined #openstack-dev | 02:47 | |
| *** shang has joined #openstack-dev | 02:53 | |
| *** fsouza has quit IRC | 03:03 | |
| andrewbogott | bsas: Devstack installs aren't persistent; as soon as you reboot your machine all is lost. If you rerun stack.sh the old state will be wiped and a new install created and started. | 03:08 |
| andrewbogott | I think there is a pathway for changing a devstack run into an actual system setup, but I haven't done it. | 03:08 |
| andrewbogott | bsas has long since left the room :( | 03:10 |
| *** thickskin has left #openstack-dev | 03:12 | |
| *** AlanClark has quit IRC | 03:15 | |
| *** fsouza has joined #openstack-dev | 03:15 | |
| winston-d | andrewbogott, what do you mean by 'all is lost'? Devstack is persistent I think. | 03:15 |
| *** lloydde has joined #openstack-dev | 03:16 | |
| *** paulstallworth has joined #openstack-dev | 03:17 | |
| *** roge has quit IRC | 03:19 | |
| *** winston-d has quit IRC | 03:22 | |
| *** lloydde has quit IRC | 03:23 | |
| *** lloydde has joined #openstack-dev | 03:24 | |
| *** littleidea has quit IRC | 03:24 | |
| *** lloydde has quit IRC | 03:24 | |
| *** Mandell has joined #openstack-dev | 03:27 | |
| paulstallworth | why might i be getting an "Invalid OpenStack Identity credentials" on keystone endpoint-list but not service-list? | 03:27 |
| paulstallworth | erm meant that for general not -dev | 03:28 |
| *** dwalleck has quit IRC | 03:30 | |
| *** winston-d has joined #openstack-dev | 03:37 | |
| *** fsouza has quit IRC | 03:37 | |
| *** kindaopsdevy has joined #openstack-dev | 03:49 | |
| *** kindaopsdevy_ has joined #openstack-dev | 03:50 | |
| *** kindaopsdevy has quit IRC | 03:53 | |
| *** kindaopsdevy_ is now known as kindaopsdevy | 03:53 | |
| *** sleepsonzzz is now known as sleepsonthefloor | 04:02 | |
| *** lloydde has joined #openstack-dev | 04:22 | |
| *** littleidea has joined #openstack-dev | 04:25 | |
| *** adjohn has quit IRC | 04:25 | |
| *** paulstallworth has quit IRC | 04:26 | |
| *** littleidea has quit IRC | 04:33 | |
| *** littleidea has joined #openstack-dev | 04:35 | |
| *** dtroyer is now known as dtroyer_zzz | 04:37 | |
| *** mdrnstm has quit IRC | 04:40 | |
| *** sniperd has quit IRC | 04:43 | |
| *** sniperd has joined #openstack-dev | 04:43 | |
| *** asalkeld has quit IRC | 04:45 | |
| *** akscram has quit IRC | 04:45 | |
| *** asalkeld has joined #openstack-dev | 04:45 | |
| *** hattwick has quit IRC | 05:02 | |
| *** RobertLaptop has joined #openstack-dev | 05:04 | |
| *** RobertLaptop has left #openstack-dev | 05:04 | |
| *** mnewby has quit IRC | 05:08 | |
| *** dhellmann has joined #openstack-dev | 05:14 | |
| *** dhellmann_ has quit IRC | 05:14 | |
| *** mattray has quit IRC | 05:14 | |
| *** zaitcev has quit IRC | 05:16 | |
| *** seats has quit IRC | 05:18 | |
| *** shiju has joined #openstack-dev | 05:19 | |
| *** kindaopsdevy has quit IRC | 05:21 | |
| *** btorch_ has joined #openstack-dev | 05:24 | |
| *** btorch_ has quit IRC | 05:31 | |
| *** bepernoot has joined #openstack-dev | 05:34 | |
| *** nati has quit IRC | 05:36 | |
| *** vincentricci has joined #openstack-dev | 05:39 | |
| *** dwalleck has joined #openstack-dev | 05:39 | |
| *** mnewby has joined #openstack-dev | 05:46 | |
| *** mnewby_ has joined #openstack-dev | 05:47 | |
| *** bepernoot has quit IRC | 05:48 | |
| *** mnewby has quit IRC | 05:51 | |
| *** mnewby_ is now known as mnewby | 05:51 | |
| *** koolhead11 has joined #openstack-dev | 05:54 | |
| *** littleidea has quit IRC | 05:57 | |
| *** lloydde has quit IRC | 06:00 | |
| *** lloydde has joined #openstack-dev | 06:01 | |
| *** lloydde has quit IRC | 06:01 | |
| *** dwalleck has quit IRC | 06:10 | |
| bodepd | can someone point me to a doc that explains how to validate a nova installation? | 06:15 |
| *** dachary has quit IRC | 06:17 | |
| koolhead11 | bodepd, http://www.hastexo.com/resources/docs/installing-openstack-essex-4-ubuntu-1204-precise-pangolin see if it helps | 06:20 |
| *** Mandell_ has joined #openstack-dev | 06:24 | |
| bodepd | koolhead11: going through that now. thanks! | 06:27 |
| *** Mandell has quit IRC | 06:27 | |
| koolhead11 | bodepd, np :) | 06:28 |
| bodepd | its actually really close, but it fails complaining that it cannot find an image when I run a nova boot | 06:34 |
| bodepd | it finds the first references images, but it fails trying to find the kernel for that image | 06:34 |
| *** journeeman has joined #openstack-dev | 06:34 | |
| *** pmyers has quit IRC | 06:38 | |
| *** pmyers has joined #openstack-dev | 06:39 | |
| *** Ryan_Lane has joined #openstack-dev | 06:39 | |
| *** sleepsonthefloor is now known as sleepsonzzz | 06:43 | |
| mikal | How do I kick jenkins to see if a test failure was real? | 06:44 |
| *** akscram has joined #openstack-dev | 06:46 | |
| ttx | mikal: you get someone to reaccept it | 06:49 |
| *** pmyers has quit IRC | 06:50 | |
| *** pmyers has joined #openstack-dev | 06:51 | |
| *** mnewby has quit IRC | 06:54 | |
| *** mnewby has joined #openstack-dev | 06:54 | |
| *** shang has quit IRC | 06:58 | |
| *** Mkenneth has joined #openstack-dev | 07:04 | |
| *** bepernoot has joined #openstack-dev | 07:06 | |
| *** devcamcar has quit IRC | 07:08 | |
| *** devcamcar has joined #openstack-dev | 07:08 | |
| *** shang has joined #openstack-dev | 07:11 | |
| *** shang has quit IRC | 07:16 | |
| mikal | Reaccept? You mean reapprove? | 07:17 |
| mikal | https://review.openstack.org/#change,6341 would apparently love a new approval then. | 07:17 |
| *** dachary has joined #openstack-dev | 07:19 | |
| *** shang has joined #openstack-dev | 07:19 | |
| *** shiju has quit IRC | 07:30 | |
| *** pixelbeat has joined #openstack-dev | 07:45 | |
| *** corXi has joined #openstack-dev | 07:51 | |
| *** hattwick has joined #openstack-dev | 07:59 | |
| *** rbasak has joined #openstack-dev | 07:59 | |
| *** fc__ has joined #openstack-dev | 08:03 | |
| *** darraghb has joined #openstack-dev | 08:05 | |
| *** derekh has joined #openstack-dev | 08:12 | |
| *** hashar has joined #openstack-dev | 08:16 | |
| *** Mandell_ has quit IRC | 08:32 | |
| *** vishy has quit IRC | 08:32 | |
| *** hugokuo has joined #openstack-dev | 08:33 | |
| *** vishy has joined #openstack-dev | 08:34 | |
| *** berendt has joined #openstack-dev | 08:51 | |
| *** eglynn has joined #openstack-dev | 08:59 | |
| *** vincentricci has quit IRC | 09:01 | |
| *** paulormg has joined #openstack-dev | 09:08 | |
| *** winston-d has quit IRC | 09:13 | |
| *** vivek has joined #openstack-dev | 09:46 | |
| *** LinuxJedi has quit IRC | 09:58 | |
| *** LinuxJedi has joined #openstack-dev | 10:02 | |
| *** LinuxJedi has quit IRC | 10:08 | |
| *** LinuxJedi has joined #openstack-dev | 10:08 | |
| *** mjfork has joined #openstack-dev | 10:12 | |
| *** shang has quit IRC | 10:39 | |
| *** hashar has quit IRC | 10:45 | |
| *** bsza has joined #openstack-dev | 10:47 | |
| *** bepernoot has quit IRC | 10:48 | |
| *** bepernoot has joined #openstack-dev | 10:49 | |
| *** mnewby has quit IRC | 10:57 | |
| *** roge has joined #openstack-dev | 11:05 | |
| *** roge has quit IRC | 11:14 | |
| *** milner has joined #openstack-dev | 11:15 | |
| sandywalsh | ttx how do I get the status on https://review.openstack.org/#change,5399 back from "abandoned" ... I was waiting for Folsom to start. | 11:27 |
| mikal | Can a nova core please kick https://review.openstack.org/#change,6341 for me? I think jenkins is a liar. | 11:27 |
| mikal | And all I want for Christmas is for someone with a strong digestion to review https://review.openstack.org/#change,6354 please | 11:29 |
| *** philr has joined #openstack-dev | 11:31 | |
| *** sandywalsh has quit IRC | 11:31 | |
| philr | Hi I have a keystone question I can't find reference to ssl configuration on Essex for port 5000 connections | 11:33 |
| *** tryggvil_ has joined #openstack-dev | 11:43 | |
| *** rods has joined #openstack-dev | 11:43 | |
| *** sandywalsh has joined #openstack-dev | 11:45 | |
| *** tryggvil_ has quit IRC | 11:47 | |
| *** markvoelker has joined #openstack-dev | 11:48 | |
| *** alaski has joined #openstack-dev | 11:52 | |
| *** philr has quit IRC | 11:55 | |
| *** shiju has joined #openstack-dev | 12:04 | |
| *** pmyers has quit IRC | 12:08 | |
| *** pmyers has joined #openstack-dev | 12:08 | |
| *** pmyers has joined #openstack-dev | 12:11 | |
| shiju | Hi, any idea on how to put 'v1.1/<tenant_id>/os-floating-ip-dns/<domain>/entries?ip=<ip>' in .wadl ? | 12:12 |
| *** hashar has joined #openstack-dev | 12:16 | |
| *** user has joined #openstack-dev | 12:26 | |
| *** user has quit IRC | 12:26 | |
| *** armaan has joined #openstack-dev | 12:42 | |
| *** mattray has joined #openstack-dev | 12:46 | |
| *** roge has joined #openstack-dev | 12:46 | |
| ttx | comstud: ping | 12:48 |
| sandywalsh | ttx how do I get the status on https://review.openstack.org/#change,5399 back from "abandoned" ... I was waiting for Folsom to start. | 12:49 |
| ttx | sandywalsh: you should have some button to resurrect it | 12:50 |
| ttx | sandywalsh: what do you have where "review" usually stands ? | 12:50 |
| sandywalsh | ttx, review/restore change/diff | 12:50 |
| ttx | let me try +0ing it | 12:51 |
| sandywalsh | ttx tried to push the branch again, tried a review, no joy | 12:51 |
| ttx | hmm, looks like my Lazarus spell needs some polish | 12:51 |
| ttx | sandywalsh: ask mtaylor/jeblair | 12:52 |
| sandywalsh | I say that every morning | 12:52 |
| sandywalsh | k ... thanks | 12:52 |
| ttx | they told me it was pretty simple, but I don't remember. | 12:52 |
| *** fsouza has joined #openstack-dev | 12:52 | |
| sandywalsh | mtaylor, jeblair: how do I get the status on https://review.openstack.org/#change,5399 back from "abandoned" ... I was waiting for Folsom to start. | 12:52 |
| ttx | something like "the submitter can resurrect it", but apparently not | 12:52 |
| ttx | sandywalsh: pushing the branch again with the changeid edited out should probably do it too | 12:53 |
| sandywalsh | but then I'll lose the review history no? | 12:53 |
| ttx | sandywalsh: unless gerrit is conscious. | 12:53 |
| ttx | sandywalsh: indeed. | 12:53 |
| sandywalsh | hmm | 12:53 |
| *** armaan has left #openstack-dev | 12:54 | |
| ttx | sandywalsh: wait for the Ci pair, I think they have the power to revive the dead. | 12:54 |
| sandywalsh | ttx agreed | 12:54 |
| sandywalsh | ttx, see ya in a week! | 12:54 |
| ttx | cheers | 12:54 |
| *** AlanClark has joined #openstack-dev | 12:57 | |
| *** stuntmachine has joined #openstack-dev | 12:57 | |
| *** littleidea has joined #openstack-dev | 13:00 | |
| *** fsouza has quit IRC | 13:04 | |
| comstud | ttx: pong | 13:05 |
| ttx | have 5 min. to discuss the Nova scaling track ? | 13:05 |
| ttx | comstud: ^ | 13:05 |
| comstud | yeah | 13:06 |
| ttx | ok, see pm | 13:06 |
| comstud | it's 6am, so i'm only somewhat conscious | 13:06 |
| comstud | :) | 13:06 |
| *** dprince has joined #openstack-dev | 13:07 | |
| *** crobinso has joined #openstack-dev | 13:08 | |
| *** dneary has joined #openstack-dev | 13:09 | |
| *** dneary has quit IRC | 13:09 | |
| *** dneary has joined #openstack-dev | 13:09 | |
| *** vincentricci has joined #openstack-dev | 13:10 | |
| *** vincentricci has quit IRC | 13:10 | |
| *** tryggvil has joined #openstack-dev | 13:14 | |
| *** oneiroi has joined #openstack-dev | 13:20 | |
| *** ayoung has joined #openstack-dev | 13:25 | |
| *** dtroyer_zzz is now known as dtroyer | 13:28 | |
| *** roge has quit IRC | 13:32 | |
| *** roge has joined #openstack-dev | 13:34 | |
| *** rgoodwin_away has joined #openstack-dev | 13:36 | |
| *** dneary has quit IRC | 13:38 | |
| *** rgoodwin has quit IRC | 13:38 | |
| *** rgoodwin_away is now known as rgoodwin | 13:38 | |
| *** slyphon has quit IRC | 13:43 | |
| *** kbringard has joined #openstack-dev | 13:50 | |
| *** dneary has joined #openstack-dev | 13:52 | |
| *** dneary has quit IRC | 13:53 | |
| *** dneary has joined #openstack-dev | 13:53 | |
| *** ayoung has quit IRC | 13:55 | |
| *** slyphon has joined #openstack-dev | 13:57 | |
| *** paulstallworth has joined #openstack-dev | 14:00 | |
| *** dneary has quit IRC | 14:01 | |
| *** GheRivero has joined #openstack-dev | 14:03 | |
| *** blamar has joined #openstack-dev | 14:10 | |
| *** shiju has quit IRC | 14:10 | |
| *** ayoung has joined #openstack-dev | 14:12 | |
| *** slyphon has quit IRC | 14:16 | |
| *** Mandell has joined #openstack-dev | 14:17 | |
| *** dneary has joined #openstack-dev | 14:18 | |
| *** dtroyer is now known as dtroyer_zzz | 14:23 | |
| *** jog0 has joined #openstack-dev | 14:25 | |
| *** journeeman has quit IRC | 14:26 | |
| *** winston-d has joined #openstack-dev | 14:27 | |
| *** lloydde has joined #openstack-dev | 14:28 | |
| *** deshantm has joined #openstack-dev | 14:28 | |
| *** dneary has quit IRC | 14:30 | |
| *** blamar has quit IRC | 14:30 | |
| *** blamar has joined #openstack-dev | 14:33 | |
| *** dwalleck has joined #openstack-dev | 14:35 | |
| *** lloydde has quit IRC | 14:42 | |
| *** lloydde has joined #openstack-dev | 14:42 | |
| *** GheRivero has quit IRC | 14:44 | |
| *** dneary has joined #openstack-dev | 14:45 | |
| *** rnirmal has joined #openstack-dev | 14:50 | |
| jeblair | sandywalsh: click "restore change" | 14:52 |
| sandywalsh | jeblair, that doesn't unwind a commit? | 14:53 |
| winston-d | jaypipes : ping | 14:53 |
| jeblair | nope, you're probably thinking of 'revert', and that's only available after it's merged | 14:53 |
| jaypipes | winston-d: pong! | 14:53 |
| *** lloydde has quit IRC | 14:53 | |
| sandywalsh | jeblair, cool, thanks. I was afraid to try that button. | 14:54 |
| jeblair | no problem! | 14:54 |
| jaypipes | sandywalsh: don't be afraid of the Big Red Button, my friend. | 14:54 |
| *** slyphon has joined #openstack-dev | 14:54 | |
| * jaypipes brings out Ren and Stimpy reference... yeah, that's right. | 14:54 | |
| winston-d | jaypipes : :) I have some questions about glance and tempest. | 14:55 |
| * sandywalsh is afraid of jaypipes let loose again in San Fran | 14:55 | |
| jaypipes | winston-d: fire away | 14:55 |
| jeblair | mikal: you should also be able to retrigger a test in jenkins | 14:55 |
| jaypipes | sandywalsh: lol :) | 14:55 |
| *** Mandell has quit IRC | 14:55 | |
| *** deshantm_ has joined #openstack-dev | 14:57 | |
| *** lorin1 has joined #openstack-dev | 14:57 | |
| *** deshantm has quit IRC | 14:58 | |
| *** jog0 has quit IRC | 14:59 | |
| jaypipes | winston-d: did you have a question for me? | 14:59 |
| *** deshantm_ is now known as deshantm | 14:59 | |
| winston-d | jaypipes : first, about glance. I tried the latest glance (trunk), set worker > 1, then started glance-api with log output to stdout. when i stopped it with ctrl-c, glance-api would end-up in a endless loop, keep poping error like 'Bad file descriptor'. this symptom is exactly the same as my multiprocess version keystone. | 15:00 |
| winston-d | jaypipes : have you seen that before? | 15:01 |
| jaypipes | winston-d: I have not :( | 15:01 |
| winston-d | jaypipes : let me see if i can reproduce it on another server. | 15:01 |
| jaypipes | winston-d: but then again, I usually start it with the glance-control executable | 15:01 |
| jaypipes | winston-d: which manages the daemon process group | 15:02 |
| *** kindaopsdevy has joined #openstack-dev | 15:02 | |
| jaypipes | winston-d: that said, if it's reproducible then we need to fix it! | 15:02 |
| winston-d | jaypipes : well, glance-control works well. | 15:02 |
| *** dwalleck_ has joined #openstack-dev | 15:04 | |
| *** dolphm has joined #openstack-dev | 15:05 | |
| *** dwalleck has quit IRC | 15:06 | |
| *** bsza has quit IRC | 15:07 | |
| *** hashar has quit IRC | 15:07 | |
| *** davlaps has joined #openstack-dev | 15:09 | |
| winston-d | jaypipes : yes, i reproduced the 'bug' | 15:09 |
| winston-d | jaypipes : the output when i press ctrl-c is like: http://paste.openstack.org/show/12630/ | 15:11 |
| winston-d | jaypipes : same issue is happening to the multiprocess Nova API and Keystone too. | 15:13 |
| jaypipes | winston-d: got it. please do file a bug. we'll get right on that. | 15:13 |
| jaypipes | winston-d: Ahmad brought up a similar issue when chatting with him offline about it. Hopefully he will add some commentary to your review | 15:13 |
| jaypipes | winston-d: my guess is that this has been a hidden issue since Swift (and Glance, which has very similar code) uses the swift-init (glance-control) daemon wrapper | 15:14 |
| *** dolphm has quit IRC | 15:15 | |
| winston-d | jaypipes : yeah, i guess so. i don't have a swift to test right now. but i did look over Swift source code, didn't find any special treatment to avoid such issue. | 15:15 |
| jaypipes | winston-d: I wish more HPers hung out on IRC... going to try and grab Ahmad to come and chat about this... bear with me. | 15:16 |
| jeblair | anotherjesse_zz: https://jenkins.openstack.org/job/gate-integration-tests-devstack-vm/3365/console | 15:16 |
| winston-d | jaypipes : sure. i'd fire a bug first. | 15:17 |
| jeblair | anotherjesse_zz: that failure is interesting; i logged into the host, and it appears that the screen window for g-api was never created, even though the console log has "screen -S stack -X screen -t g-api" | 15:17 |
| jeblair | anotherjesse_zz: have you ever seen an error like that? | 15:17 |
| jeblair | anotherjesse_zz: (you can also see that the g-api screen log file also wasn't created, which i think helps confirm that command didn't work) | 15:18 |
| *** zaitcev has joined #openstack-dev | 15:18 | |
| *** negronjl has quit IRC | 15:19 | |
| jaypipes | winston-d: cool. sent an email to ahmad and phil day to hop on IRC. | 15:19 |
| *** nati has joined #openstack-dev | 15:20 | |
| *** dolphm has joined #openstack-dev | 15:23 | |
| *** reed has joined #openstack-dev | 15:24 | |
| *** kindaopsdevy has left #openstack-dev | 15:24 | |
| *** dwalleck_ has quit IRC | 15:25 | |
| *** lloydde has joined #openstack-dev | 15:25 | |
| mtaylor | jeblair: jeez - we have to account for _screen_ failures now? | 15:25 |
| *** dwalleck has joined #openstack-dev | 15:26 | |
| *** bepernoot has quit IRC | 15:27 | |
| *** Ahmad_HP has joined #openstack-dev | 15:29 | |
| sdague | jeblair: you didn't happen to fill up the disk or anything did you? | 15:29 |
| Ahmad_HP | Hi Jay, | 15:29 |
| Ahmad_HP | sorry just check your email | 15:29 |
| jeblair | /dev/vda 9.9G 2.0G 7.4G 22% / | 15:30 |
| winston-d | :) | 15:30 |
| jeblair | sdague: nope, good idea though. i'd expect to see some kind of error message in the console log though. afaict, there was none, and the screen command exited with success. | 15:30 |
| *** dolphm has quit IRC | 15:31 | |
| sdague | yeh, that screen fail seems really odd, though I guess if the glance command didn't work at all, it may not have created the log file. I kind of think the screen fail is a red herring | 15:32 |
| jaypipes | Ahmad_HP: hi! | 15:32 |
| jaypipes | winston-d: Ahmad_HP is here :) | 15:32 |
| winston-d | hi Ahmad_HP | 15:32 |
| jaypipes | Ahmad_HP: thx for popping on IRC. so the reason we are comparing with Glance is because the code Huang proposed for merging is very similar to Glance+Swift | 15:33 |
| *** spiffxp has joined #openstack-dev | 15:34 | |
| jeblair | sdague: stack.sh creates a screen window, turns on logging, then runs the glance command. i'm pretty sure creating the window failed because (1) it wasn't there when i logged into the host and resumed the screen, and (2) the log file was never created. i think if the window wasn't created properly, that explains why there was no log file, and g-api didn't start. it could be that it was created and then closed, but i see no reason why | 15:35 |
| Ahmad_HP | No problem Jay. I am glad that @winton-d has noticed this thing early | 15:35 |
| Ahmad_HP | I think we slightly differ from glance... | 15:35 |
| Ahmad_HP | because we have a Grand parent - Parent - CHILD WORKERS model | 15:36 |
| Ahmad_HP | where as glance would have Parent - Child works model | 15:36 |
| *** ecarlin has joined #openstack-dev | 15:36 | |
| *** ecarlin has quit IRC | 15:36 | |
| jaypipes | Ahmad_HP: yes, that sounds correct | 15:37 |
| jaypipes | winston-d: when Ahmad says "we", he is referring to HPCS' internal Nova branch that has multiprocessing nova-api... | 15:37 |
| *** negronjl has joined #openstack-dev | 15:37 | |
| Ahmad_HP | @winston-d, can you please point me to your feature branch on gerrit please? | 15:37 |
| jaypipes | Ahmad_HP: are you familiar with the glance-control daemon control script? | 15:37 |
| *** ayoung has quit IRC | 15:38 | |
| Ahmad_HP | No Jay, I have very limited knowledge of glance code base, But I have backported quite a few things from glance i.e. multiprocessing concept and sqlalchemy db api sessions handling | 15:38 |
| jaypipes | Ahmad_HP: https://review.openstack.org/#change,5762 | 15:39 |
| Ahmad_HP | Thanks Jay | 15:39 |
| winston-d | jaypipes, Ahmad_HP yes, the reason why I added one grand parent (aka master_worker in the code) is to be able to pass the unittest, because if parent-child worker mode like glance, the stop() method of WSGI service would not only kill the service but also the unittest itself by 'os.killpg(0...)' | 15:39 |
| jaypipes | Ahmad_HP: no prob. the glance-control daemon wrapper is essentially what does the Grandparent -> Parent -> Children model in Glance. | 15:39 |
| jaypipes | Ahmad_HP: but if you start glance-api directly (without using glance-control wrapper), you get the hanging file descriptors... | 15:40 |
| *** issackelly has joined #openstack-dev | 15:40 | |
| jaypipes | Ahmad_HP: when you SIGTERM the glance-api process. | 15:40 |
| Ahmad_HP | so yes. in @winston-d branch, The grand parent is the main process which spawns the service............It then invokes start method in WSGI class......which creates the master process. The master process then create workers | 15:40 |
| Ahmad_HP | so it has 3 layers | 15:40 |
| jaypipes | right. | 15:40 |
| winston-d | correct | 15:41 |
| jaypipes | Ahmad_HP: for reference... here is Glance server forking code: https://github.com/openstack/glance/blob/master/glance/common/wsgi.py | 15:42 |
| *** dolphm has joined #openstack-dev | 15:42 | |
| Ahmad_HP | When I backported the change from Glance in to HP nova, then we had this serious problem of Zombi processes left over, especially when CHEF deploys the services | 15:42 |
| *** markmc has joined #openstack-dev | 15:42 | |
| jaypipes | Ahmad_HP: gotcha. and your solution had to do with manually catching SIGTERM/SIGINT inside the parent... is that correct? | 15:42 |
| Ahmad_HP | There might be a clean and quick way to address this problem, but I implemented that resilient layer in the following way: | 15:42 |
| winston-d | Ahmad_HP : i've seen similar issue too. | 15:43 |
| Ahmad_HP | When any parent proecess gets killed, then it should clear out all the child process groups and it's own process group | 15:43 |
| Ahmad_HP | If child detects that its parent has been died, then it kills itself | 15:43 |
| Ahmad_HP | yeah right, Jay | 15:44 |
| openstackgerrit | Verification of a change to openstack/nova failed: migration_get_all_unconfirmed() now uses lowercase "finished" Fixes bug 977719 https://review.openstack.org/6402 | 15:44 |
| uvirtbot | Launchpad bug 977719 in nova "migration_get_all_unconfirmed() should use lower case "finished"" [Undecided,In progress] https://launchpad.net/bugs/977719 | 15:44 |
| Ahmad_HP | @winston-d. Please verify your code both with kill -9 and kill -15 | 15:44 |
| jaypipes | Ahmad_HP: See this line here: https://github.com/openstack/glance/blob/master/glance/common/wsgi.py#L153 | 15:45 |
| Ahmad_HP | In openstack, we want to recover the whole api if the SIGKILL or SIGTERM was raised on any process with in whole process group | 15:45 |
| jaypipes | Ahmad_HP: are you suggesting that the call in that block to killpg(0, ...) is at fault? | 15:45 |
| jaypipes | Ahmad_HP: that instead the server should signal to kill each child separastely? | 15:45 |
| Ahmad_HP | killpg(0, ...) is fine. 0 means killing the whole process group..................which kills the children as well | 15:46 |
| annegentle | dolphm: around? | 15:46 |
| dolphm | annegentle: yes | 15:46 |
| jaypipes | Ahmad_HP: ah! I see now... in the child process itself, the child needs to check the parent is killed... | 15:46 |
| Ahmad_HP | Actually, when CHEF kills the process, according to my understanding it only send the SIGTERM to main process... | 15:47 |
| annegentle | dolphm: do you have a minute to look at bug 973477 with me? | 15:47 |
| uvirtbot | Launchpad bug 973477 in openstack-manuals "Difficulties with keystone docs with mysql" [Medium,Confirmed] https://launchpad.net/bugs/973477 | 15:47 |
| jaypipes | Ahmad_HP: well that *seems* to be handled correctly in the current code, no? SIGTERM is caught in the parent and in each child AFAICT... | 15:47 |
| annegentle | dolphm: I've got a VM running where I'm still getting an error, Access denied for user 'keystone'@'localhost' (using password: YES) | 15:48 |
| dolphm | annegentle: sure, reading | 15:48 |
| Ahmad_HP | The use of sys.exit(1) is very bad.... I was using this in my code intially and this lead us to ZOMBIE process scnarios 'rarely' | 15:48 |
| jaypipes | Ahmad_HP: or perhaps you are saying that this block: https://github.com/openstack/glance/blob/master/glance/common/wsgi.py#L215 also needs a SIGTERM (non-default) handler? | 15:48 |
| annegentle | dolphm: it has to be something wrong with either the GRANT or that mysql isn't configured for localhost? | 15:48 |
| Ahmad_HP | we got rid of sys.exit from our HP nova implementation | 15:48 |
| mtaylor | Ahmad_HP: what do you use instead of sys.exit? | 15:49 |
| dolphm | annegentle: what's the result of the two commands i suggested? | 15:49 |
| jaypipes | Ahmad_HP: ok. so here: https://github.com/openstack/glance/blob/master/glance/common/wsgi.py#L197 we should just let the KeyboardInterrupt block exit normallY? | 15:49 |
| dolphm | annegentle: that should verify if grant is working correctly | 15:49 |
| jaypipes | Ahmad_HP: or should we return 1 or something? | 15:49 |
| dolphm | annegentle: and that the user issuing the grant has *permission* to issue the grant | 15:49 |
| Ahmad_HP | @mtaylor: We didn't use sys.exit for killing the process becaue it has child process tree underneath | 15:50 |
| Ahmad_HP | instead, we raise SIGTERM on all the process group and let the process gets stopped cleanly | 15:50 |
| jaypipes | Ahmad_HP: understood. OK, I'll make some quick changes to my local branch and test if removing the sys.exit -- and re-raising the KeyboardInterrupt -- solves the issue. | 15:51 |
| annegentle | dolphm: ok, in the Host column for the User with keystone, it shows % | 15:51 |
| annegentle | dolphm: so there must be something wrong with the GRANT command? | 15:51 |
| dolphm | annegentle: i think that's fine -- it means 'any host' | 15:51 |
| annegentle | dolphm: GRANT ALL ON keystone.* TO 'keystone'@'%' IDENTIFIED BY 'yourpassword'; | 15:52 |
| annegentle | dolphm: ok | 15:52 |
| dolphm | annegentle: what's the error from db_sync? is that the access denied for user 'keystone'@'localhost'? | 15:52 |
| Ahmad_HP | @jaypiples Does SIGTERM raises that keyboardinterrupt exception? I doubt it would? | 15:52 |
| winston-d | i need a reboot to my devstack environment to verify that with my nova patch too. | 15:52 |
| annegentle | dolphm: yep | 15:52 |
| winston-d | brb | 15:52 |
| annegentle | dolphm: my my.cnf file for mysql uses bind_address 0.0.0.0 | 15:53 |
| dolphm | annegentle: try to start a mysql session with the expected credentials: mysql -u keystone -p yourpassword | 15:53 |
| sdague | is there an easy way to run just one unit test in the nova/tests/ dir? | 15:53 |
| annegentle | dolphm: same error | 15:53 |
| *** winston-d has quit IRC | 15:53 | |
| annegentle | dolphm: I can start mysql with root | 15:53 |
| annegentle | dolphm: but not with the keystone user | 15:53 |
| jaypipes | Ahmad_HP: hmm, no, I think that is SIGHUP... | 15:53 |
| jaypipes | Ahmad_HP: but who knows.. :) I'll research and play around with it. | 15:53 |
| dolphm | annegentle: as root, can you list all users with: select * from mysql.user; | 15:53 |
| annegentle | dolphm: and in the keystone user row, there's a bunch of N | 15:54 |
| dolphm | annegentle: and 'keystone' does appear? | 15:54 |
| jaypipes | Ahmad_HP: thx very much for hopping on to discuss! Very much appreciated! | 15:54 |
| dolphm | N? | 15:54 |
| annegentle | dolphm: here you go: http://paste.openstack.org/show/12631/ | 15:54 |
| Ahmad_HP | Ok @jaypipes. Having said that, I would suggest that we test this rigorously. In our HP source, I tested with both kill -9 and kill -15 and everything was working fine... | 15:54 |
| *** winston-d has joined #openstack-dev | 15:54 | |
| annegentle | dolphm: it's as if that GRANT command doesn't actually grant | 15:54 |
| annegentle | dolphm: and I'm not sure how to fix it | 15:55 |
| Ahmad_HP | but when we deployed through chef then occassionaly we weree seeing the zombies | 15:55 |
| Ahmad_HP | so then I debugged further and removed sys.exits. | 15:55 |
| Ahmad_HP | until now, it works awesome | 15:55 |
| dolphm | annegentle: what version of mysql? | 15:55 |
| dolphm | annegentle: to make sure i'm looking at the correct docs | 15:55 |
| Ahmad_HP | 'Zombie' was pretty famous in our QA team last month :) | 15:56 |
| *** ayoung has joined #openstack-dev | 15:56 | |
| jaypipes | Ahmad_HP: rock on. :) I will do my best to bring your knowledge into Glance and winston-d | 15:56 |
| jaypipes | 's patches. | 15:56 |
| jaypipes | Ahmad_HP: ty! | 15:56 |
| *** armaan has joined #openstack-dev | 15:56 | |
| annegentle | dolphm: Ver 14.14 Distrib 5.5.22 | 15:57 |
| Ahmad_HP | @jaypipes. thanks | 15:57 |
| jaypipes | Ahmad_HP: of course, if you could wave your magic wand and get me 8 more hours in a day, that would be awesome too! ;) | 15:57 |
| jaypipes | Ahmad_HP: or ask QA to do that ;) | 15:57 |
| Ahmad_HP | haha | 15:57 |
| dolphm | annegentle: ALL is definitely valid then, according to http://dev.mysql.com/doc/refman/5.5/en/grant.html | 15:58 |
| Ahmad_HP | 'Zombies' were really annoying in our case, because they used to appear randomly and very rarely | 15:58 |
| winston-d | :) | 15:58 |
| annegentle | dolphm: I also know this command to work when using Ubuntu 11.10 base. | 15:59 |
| annegentle | dolphm: hm. | 15:59 |
| Ahmad_HP | and we changed init.d script to respawn the service, so if zombies appear then service respawns every 3 minute in our case and eventually all the cloud goes does :) | 15:59 |
| annegentle | dolphm: well I need to head to a meeting but wanted to see if I can troubleshoot this further. | 16:00 |
| annegentle | dolphm: it's quite a mystery to me right now | 16:00 |
| *** ghe has joined #openstack-dev | 16:01 | |
| *** ghe is now known as Guest25274 | 16:01 | |
| jaypipes | Ahmad_HP: but since you refactored the daemonization stuff, the zombies have disappeared, correct? | 16:02 |
| *** dolphm has quit IRC | 16:02 | |
| Ahmad_HP | we don't see any zombies since last 3 weeks.....things looking good | 16:02 |
| jaypipes | awesome. | 16:03 |
| *** ayoung has quit IRC | 16:03 | |
| Ahmad_HP | and we have around 2000 nodes | 16:03 |
| *** rkukura has quit IRC | 16:04 | |
| Ahmad_HP | infact 3k I guess | 16:04 |
| winston-d | 2000 nodes for testing? impressive! | 16:04 |
| jaypipes | Ahmad_HP: that's all? ;) | 16:04 |
| openstackgerrit | Verification of a change to openstack/nova failed: Standardize quota flag format https://review.openstack.org/6406 | 16:04 |
| *** andrewsben_zz is now known as andrewsben | 16:04 | |
| jaypipes | winston-d: production, not testing :) | 16:04 |
| jaypipes | winston-d: HP Cloud :) | 16:04 |
| *** LinuxJedi has quit IRC | 16:04 | |
| winston-d | jaypipes, that's not trycloud.org, right? | 16:04 |
| jaypipes | winston-d: now... if ONLY we could have a nice little 2K node cluster for TryStack! :) | 16:05 |
| jaypipes | winston-d: no, definitely not! | 16:05 |
| *** LinuxJedi has joined #openstack-dev | 16:05 | |
| Ahmad_HP | not for testing only. For various things i.e. staging and testing. But yeah testing nodes are a lot | 16:05 |
| jaypipes | winston-d: we have 60 nodes provisioned for TryStack, though. Which is still excellent. | 16:05 |
| jaypipes | winston-d: 60 nodes across 2 zones -- HP and Dell. | 16:05 |
| *** ncode has joined #openstack-dev | 16:06 | |
| *** ayoung has joined #openstack-dev | 16:06 | |
| winston-d | i see. we only have like ~40 nodes. | 16:06 |
| Ahmad_HP | we needed more nodes to test db load in real time scenarios | 16:06 |
| winston-d | Ahmad_HP, what's ratio of API nodes vs. compute nodes if it's ok to share? | 16:08 |
| winston-d | Ahmad_HP, you must also have a separate cloud to generate pressure for those 3K node production cloud. | 16:09 |
| Ahmad_HP | I don't think I can share this publicly. But it is very configurable. | 16:09 |
| winston-d | wow | 16:09 |
| Ahmad_HP | We have real clones of production systems, so we do one part of testing there... Rest of testing infrastructure is separate | 16:11 |
| winston-d | no problem. i just wonder how good can API services be if multprocess is enabled. | 16:11 |
| Ahmad_HP | Testing has got prettty matured here | 16:11 |
| winston-d | nice~ | 16:12 |
| *** eglynn has quit IRC | 16:12 | |
| Ahmad_HP | Multiprocessing works awesome. Ahhhh I remember now... | 16:12 |
| jaypipes | Ahmad_HP: testing is matured, yes, but unfortunately a lot of it is a black hole that hasn't been contributed upstream... getting better about that over time, though :) | 16:13 |
| Ahmad_HP | @Jay: As you introduce multiprocessing, we got issue with security groups. Because API routes requests very fast and then compute managers fails to handle that many.... | 16:13 |
| Ahmad_HP | so we had to fix that as well | 16:13 |
| *** Mandell has joined #openstack-dev | 16:13 | |
| Ahmad_HP | but you might not see that in essex | 16:13 |
| Ahmad_HP | 'might' | 16:13 |
| jaypipes | Ahmad_HP: is this a race condition you are describing? Where security groups are tried to be created before, say, other work in instance launch has completed? | 16:14 |
| Ahmad_HP | @jaypipes. Yeah agree. It bacame really hard to push some things because essex moved away from diablo massively | 16:15 |
| jaypipes | winston-d: note that I added a comment on your code review about needing to update the random seed generation for generate_uid() in nova/utils.py. Ahmad_HP ran into that one too :) | 16:15 |
| Ahmad_HP | I am not sure about this security bug thing. It was a bug before, but it was not obvious.......When API got faster, it bacame obvious | 16:15 |
| Ahmad_HP | I will double check with the guy who worked on that | 16:15 |
| winston-d | jaypipes, yeah, i noticed that. will work on that very soon, maybe during design summit. :) | 16:16 |
| jaypipes | Ahmad_HP: cool. thx :) | 16:16 |
| winston-d | Ahmad_HP, do you guys use tempest to test your production cloud? | 16:17 |
| Ahmad_HP | I am not familiar with testing team. They might use it. | 16:17 |
| winston-d | Ahmad_HP, :) thx | 16:18 |
| *** oneiroi is now known as Oneiroi^gone | 16:21 | |
| winston-d | jaypipes, about the glance bug I just fired, is removing 'sys.exit' going to fix it? | 16:23 |
| *** dtroyer_zzz is now known as dtroyer | 16:24 | |
| jaypipes | winston-d: not sure. I'm going to test that this afternoon (or if you have time, you can do so...) I am working on a design summit proposal at this time... probably going to be another hour until I get to the testing | 16:24 |
| jaypipes | winston-d: yes, HP QA team is actively involved in the Temepst project. They've contributed about 50-60% of the Compute API tests to date. | 16:25 |
| *** Ryan_Lane has quit IRC | 16:25 | |
| winston-d | jaypipes, i can do a quick test before i fell asleep. :) | 16:25 |
| jaypipes | winston-d: ok, that would be great. thx! | 16:25 |
| *** ayoung has quit IRC | 16:26 | |
| *** dneary has quit IRC | 16:29 | |
| *** corXi has quit IRC | 16:30 | |
| markmc | mtaylor, jeblair, hey | 16:32 |
| markmc | mtaylor, jeblair, can we get some help getting https://review.openstack.org/#change,5956 merged ? (nova security fix for diablo) | 16:32 |
| markmc | mtaylor, jeblair, some *more* help, rather :) | 16:33 |
| winston-d | jaypipes, replacing sys.exit in wait_on_children with os.killpg(0, signal.SIGTERM) is working. | 16:33 |
| jaypipes | winston-d: good news :) | 16:34 |
| *** cdub has quit IRC | 16:34 | |
| markmc | mtaylor, jeblair, looks like glance is still broken too :( https://review.openstack.org/#change,6152 | 16:34 |
| *** cdub has joined #openstack-dev | 16:35 | |
| Ahmad_HP | @winston-d Awesome | 16:35 |
| Ahmad_HP | winston-d: can you do one more check please............ | 16:36 |
| Ahmad_HP | ps axf | grep nova-api | 16:36 |
| Ahmad_HP | grab the process id of process starting with 'su' | 16:36 |
| Ahmad_HP | then kill -9 and kill -15 on that | 16:37 |
| winston-d | Ahmad_HP, ah, i was looking at glance-api. but i guess you also want to apply same check to glance too? | 16:37 |
| Ahmad_HP | yes definitely | 16:38 |
| jaypipes | winston-d: glance-registry, not glance :) glance is the CLI tool :) | 16:38 |
| *** vincentricci has joined #openstack-dev | 16:39 | |
| winston-d | Ahmad_HP, hmm, i see zombie processes. | 16:40 |
| *** armaan has left #openstack-dev | 16:41 | |
| Ahmad_HP | so this means, you need to put code in the child to check the state of the parent | 16:41 |
| Ahmad_HP | That's how we did in HP source | 16:41 |
| winston-d | the SIGTERM handler in parent process should have done killpg(0, SIGTERM). why doesn't it work? | 16:41 |
| *** maplebed has joined #openstack-dev | 16:46 | |
| smoser | vishy, ping | 16:50 |
| smoser | bug 977765 | 16:51 |
| uvirtbot | Launchpad bug 977765 in nova "Image registration (ec2) is broken using deprecated auth" [High,New] https://launchpad.net/bugs/977765 | 16:51 |
| smoser | we are fully aware of the string "deprecated" there, but my feeling is that "deprecated" and "will break 2 weeks before release" are different things. | 16:51 |
| *** heckj has joined #openstack-dev | 16:52 | |
| smoser | so i'm asking if you or someone else could take a look at what adam_g came up with as to whether its reasonable, or suggest something more-so | 16:52 |
| vishy | smoser: sup? | 16:52 |
| *** nati has quit IRC | 16:52 | |
| *** ayoung has joined #openstack-dev | 16:53 | |
| vishy | smoser: is glance deployed without auth? | 16:53 |
| smoser | hm.. probably. | 16:54 |
| smoser | adam_g, knows more, and reproduced on devstack. | 16:54 |
| vishy | smoser: fyi, deprecated auth is not supported in essex | 16:54 |
| vishy | we only left it in to allow people to migrate | 16:54 |
| smoser | but, eyah, we're using deprecated auth, and i do not believe glance is exposed api wise at all. | 16:54 |
| smoser | vishy, right, but you didn't leave it | 16:54 |
| smoser | :) | 16:54 |
| smoser | its not there right now to migrate. | 16:55 |
| vishy | smoser: the plan is upgrade to essex, migrate | 16:55 |
| vishy | actually it is | 16:55 |
| vishy | it was supposed to be gone in essex | 16:55 |
| vishy | but we had to leave the code in to allow for migrations | 16:55 |
| *** derekh has quit IRC | 16:55 | |
| smoser | hm.. well, if it werent' functional up until 2 weeks ago, we would proably be off of it. | 16:55 |
| smoser | and i doubt in 3 months time the oepnstack installation that is using it will continue. | 16:56 |
| *** Ryan_Lane has joined #openstack-dev | 16:56 | |
| smoser | but something regressed when they upgraded last friday to release from another snapshot (of un-determined version at the moment, but reasonably recent) | 16:56 |
| smoser | i suspect that that regression was non intended. | 16:57 |
| smoser | anyway, vishy if you could take a quick gander there, we'd appreciate it. | 16:58 |
| smoser | right now, my testing grounds of many many things is broken. | 16:58 |
| *** AlanClark is now known as AlanClark_away | 16:58 | |
| vishy | smoser: yeah i know the commit that broke it | 16:58 |
| *** Guest25274 has quit IRC | 17:02 | |
| *** darraghb has quit IRC | 17:02 | |
| *** jdurgin has joined #openstack-dev | 17:04 | |
| *** slyphon is now known as slyphon_away | 17:05 | |
| *** stuntmachine has quit IRC | 17:06 | |
| vishy | smoser: a6ac8af69351cb39aa07f53e3327ff29b90383bc | 17:06 |
| bodepd | I am having trouble booting a vm with an image that refers to another image (like a disk image with a separate kernel) | 17:07 |
| vishy | smoser: with glance using no auth it doesn't have an owner field | 17:07 |
| bodepd | is this still supported? | 17:07 |
| *** shiju has joined #openstack-dev | 17:07 | |
| vishy | so it needs to switch back to properties project id | 17:07 |
| winston-d | Ahmad_HP, how do you check parent process's status? by checking ppid is 1 or prctl()? | 17:08 |
| *** heckj has quit IRC | 17:09 | |
| *** ayoung has quit IRC | 17:09 | |
| Ahmad_HP | i check through: | 17:09 |
| Ahmad_HP | try: | 17:09 |
| Ahmad_HP | os.kill(pid, 0) | 17:10 |
| Ahmad_HP | except OSError | 17:10 |
| Ahmad_HP | if it throws OSError, then it means process is dead | 17:10 |
| Ahmad_HP | otherwise process is alive | 17:10 |
| *** Mkenneth has quit IRC | 17:11 | |
| winston-d | great, thanks for the tip, i tried checking ppid==1, but seemed not work as expected. | 17:12 |
| *** cdub has quit IRC | 17:12 | |
| winston-d | Ahmad_HP, os.kill(pid, 0) 'pid' here is pid for parent? | 17:13 |
| Ahmad_HP | any process id that you want to check............. | 17:14 |
| Ahmad_HP | in that case, yes it will be parent | 17:14 |
| winston-d | but when do you do the checking? I assume child should do checking when received SIGTERM? | 17:15 |
| vishy | smoser: is this really high priority for you guys? Seems like a terrible idea to stay on deprecated auth with nova-objectstore | 17:15 |
| vishy | smoser: anyway I posted how to fix in the bug | 17:15 |
| smoser | vishy, its high priority in that its broken. i can't publish new images. | 17:16 |
| smoser | no one thinks they're going to be on deprecated auth in 3 months. | 17:17 |
| smoser | but they're there now. | 17:17 |
| vishy | smoser: but why did they upgrade then? | 17:17 |
| vishy | ! | 17:17 |
| smoser | and it was unexpected and significat breakage. | 17:17 |
| *** rohit404 has joined #openstack-dev | 17:17 | |
| smoser | if it failed on the upgrade, then we wouldn't be here. | 17:17 |
| smoser | ie, they upgraded to essex long ago. | 17:17 |
| vishy | smoser: upgrade should be install essex/migrate to keystone | 17:17 |
| vishy | smoser: essex w/o keystone == not ok | 17:18 |
| vishy | smoser: but anyway as i said in the bug report it is a pretty easy fix | 17:18 |
| smoser | k. thanks. | 17:18 |
| vishy | smoser: I suspect image owner is broken as well | 17:18 |
| vishy | so make sure to take care of both places | 17:18 |
| smoser | yeah, adam_g pointed that out. | 17:18 |
| smoser | (there was a patch on that bug) | 17:18 |
| *** adjohn has joined #openstack-dev | 17:19 | |
| *** cdub has joined #openstack-dev | 17:19 | |
| smoser | oh wait. did you mena something other than user_id and nnproject_id ? | 17:19 |
| *** mdrnstm has joined #openstack-dev | 17:20 | |
| *** bepernoot has joined #openstack-dev | 17:22 | |
| *** alaski has quit IRC | 17:22 | |
| *** nati has joined #openstack-dev | 17:23 | |
| *** markvoelker has quit IRC | 17:23 | |
| *** alaski has joined #openstack-dev | 17:24 | |
| smoser | vishy, so you think the patch there is reasonable for us to at least carry? | 17:25 |
| jaypipes | dprince: hi dan! hoping you might be able to look into the Torpedo failure on this review: https://review.openstack.org/6407 | 17:25 |
| smoser | http://paste.ubuntu.com/922918/ | 17:25 |
| jaypipes | dprince: basically, all the patch does is remove an invalid section in the glance-api-paste.ini file for cachemanage (without keystone). | 17:25 |
| vishy | smoser: a) user_id isn't necessary b) i'm not totally sure that is the best place for it, but i guess i don't mind. c) you also need to fix the display of ImageOwnerId | 17:25 |
| jaypipes | dprince: I'm wondering if the SmokeStack creates the API server with flavor = cachemanage? | 17:26 |
| smoser | vishy, ok. thanks. | 17:26 |
| jaypipes | dprince: if so, that will need to be changed to flavor = keystone+cachemanage (as cachemanage middleware requries keystone...) | 17:26 |
| vishy | smoser: i also wouldn't mind that going into stable/essex mentioning that it isn't needed in master because deprecated auth will be gone | 17:27 |
| *** andrewsben is now known as andrewsben_zz | 17:31 | |
| rohit404 | vishy: http://paste.openstack.org/show/12637/ (bad idea ?) | 17:35 |
| *** dolphm has joined #openstack-dev | 17:35 | |
| vishy | jeblair mtaylor : do we have a plan for publishing clients to pypi? | 17:35 |
| *** Ahmad_HP has quit IRC | 17:36 | |
| vishy | rohit404: It might work in vlan mode | 17:36 |
| rohit404 | rohit404: using quantum manager | 17:36 |
| rohit404 | vishy: ^ | 17:36 |
| jeblair | vishy: i believe the current process is poke mtlayor; to be revised at summit. | 17:36 |
| vishy | rohit404: hmm, that i don't know | 17:36 |
| vishy | rohit404: you can definitely run to dnsmasqs in different vlans although there are issues | 17:37 |
| dprince | jaypipes: yes. | 17:39 |
| *** armaan has joined #openstack-dev | 17:39 | |
| *** andrewsben_zz is now known as andrewsben | 17:40 | |
| dtroyer | mtaylor jeblair: Are you the guys who can make a stable/essex branch for DevStack for me? | 17:40 |
| dprince | jaypipes: I'll take a look and if that is the way we are going we can adjust the configs accordingly. | 17:40 |
| *** Oneiroi^gone is now known as oneiroi | 17:40 | |
| jaypipes | dprince: ty Dan~! | 17:41 |
| jk0 | any keystone cowboys available to check this out for me? https://review.openstack.org/#change,6387 | 17:42 |
| jk0 | also a devstack core? https://review.openstack.org/#change,6388 | 17:42 |
| jk0 | fixes all keystone unit test/devstack issues | 17:43 |
| vishy | tr3buchet: I don't think we ever finished the discussion on this one https://review.openstack.org/#change,6366 | 17:43 |
| *** koolhead11 has quit IRC | 17:44 | |
| *** dolphm has quit IRC | 17:45 | |
| *** andrewsben is now known as andrewsben_zz | 17:46 | |
| rohit404 | vishy: yeah, i have run two dnsmasq's on different vlans succesfully but they were for different subnets...i'm not clear if dnsmasq worked fine for one vm network why not for the second vm network... | 17:47 |
| *** alaski has quit IRC | 17:47 | |
| vishy | rohit404: so dnsmasq doesn't support bind to device | 17:47 |
| openstackgerrit | Verification of a change to openstack/nova failed: Improved localization testing. https://review.openstack.org/4499 | 17:47 |
| vishy | rohit404: but i would think the broadcast request would still work as long as the two devices are layer2 isolated | 17:48 |
| *** heckj has joined #openstack-dev | 17:48 | |
| rohit404 | vishy: its binding to the bridge interfaces for the two networks, which have interfaces on different vlans | 17:48 |
| rohit404 | vishy: yes, the broadcast is going thru...DHCPDISCOVERY messsage..but the DHCPOFFER doesnt get to the second vm | 17:49 |
| rohit404 | which is a unicast i think | 17:49 |
| vishy | outgoing unicast should work | 17:50 |
| vishy | so i suspect a different issue | 17:50 |
| *** GheRivero has joined #openstack-dev | 17:51 | |
| rohit404 | the dhcp gateway conf file also gets updated with the mac, ip address | 17:51 |
| rohit404 | i also tried to launch the vm's without the instance iptable rules (incase they were dropping) but that didnt help | 17:53 |
| *** lloydde has quit IRC | 17:53 | |
| *** Ryan_Lane has quit IRC | 17:54 | |
| *** Ryan_Lane has joined #openstack-dev | 17:54 | |
| *** slyphon_away is now known as slyphon | 17:55 | |
| winston-d | jaypipes, just want to confirm if my understanding is correct: child process needs to check parent process status once in a while (say, every 10 secs) to handle the case when parent is killed. | 17:56 |
| jaypipes | winston-d: exactly. | 17:56 |
| jaypipes | winston-d: one sec, lemme pull some code for you from the HPCS branch... | 17:57 |
| jk0 | c | 17:57 |
| jk0 | er, excuse me. trouble seeing with this corneal abrasion :) | 17:57 |
| winston-d | jaypipes, ok. so timer is needed. | 17:57 |
| *** mnewby has joined #openstack-dev | 17:58 | |
| tr3buchet | vishy: true | 17:58 |
| *** andrewsben_zz is now known as andrewsben | 18:03 | |
| *** Ahmad_HP has joined #openstack-dev | 18:05 | |
| *** lloydde has joined #openstack-dev | 18:06 | |
| winston-d | jaypipes, Ahmad_HP see if this works for you. http://paste.openstack.org/show/12639/ | 18:07 |
| vishy | tr3buchet: so the question was should both calls go through compute? | 18:07 |
| jaypipes | winston-d: I think that solution may be more than what is needed... gimme a few more seconds to paste an alternative. | 18:08 |
| dprince | Where do we stand on the Nova Diablo gating? I'd like this branch to land: https://review.openstack.org/#change,5956 | 18:09 |
| *** vricci has joined #openstack-dev | 18:09 | |
| dprince | As I understand it the issue is blocked because of a bad python-novaclient build that is in PyPi.... | 18:10 |
| dprince | jeblair: ^^ | 18:11 |
| dprince | Give its been a week I'd like to consider the option that we 'ungate' diablo... | 18:11 |
| *** vincentricci has quit IRC | 18:12 | |
| *** oneiroi is now known as Oneiroi^gone | 18:12 | |
| sdague | winston-d: the children aren't getting SIGCHILD? | 18:13 |
| *** vincentricci has joined #openstack-dev | 18:13 | |
| *** bepernoot has quit IRC | 18:14 | |
| *** vricci has quit IRC | 18:14 | |
| winston-d | sdague, you mean the parent aren't getting SIGCHILD? | 18:15 |
| sdague | oh, you're right, I was thinking about this in reverse. Never mind | 18:15 |
| *** novas0x2a|laptop has joined #openstack-dev | 18:16 | |
| sdague | time for afternoon coffee | 18:16 |
| rmk | Are there specifics anywhere around the requirements to get boot from volume working? | 18:16 |
| openstackgerrit | Verification of a change to openstack/keystone failed: Fix "it's" grammar errors. https://review.openstack.org/6418 | 18:17 |
| *** bepernoot has joined #openstack-dev | 18:18 | |
| tr3buchet | vishy: i'm not sure why it would need to.. | 18:21 |
| tr3buchet | vishy: this works for now until we finish the decorator | 18:21 |
| *** shiju has quit IRC | 18:21 | |
| tr3buchet | vishy: once the decorator is upstream, we can just decorate network_api.associate_floating_ip(...) and it will automatically update the cache | 18:22 |
| tr3buchet | without having to call get_instance_nw_info elsewhere | 18:22 |
| *** bepernoot has quit IRC | 18:23 | |
| novas0x2a|laptop | does anyone have an opinion on presence or absence of the "Conflicts" block when backporting a patch? (i.e. leave it in? take it out? i usually leave it in, jerdfelt prefers not, and i think I've seen both from others) | 18:25 |
| jeblair | dprince: alternately, we could fix the underlying problem. stable/diablo nova depends on an unversioned python-novaclient. it could be updated to depend on a particular version. and if that particular version is broken, it could be updated to be fixed. | 18:27 |
| dprince | jeblair: I'd love to do that. Is there anything I can do to help? | 18:27 |
| winston-d | jaypipes, i need to get some sleep. will check your solution when i wake up... bye | 18:28 |
| dprince | jeblair: The git tag for python-novaclient 2.6.0 doesn't even seem to exist? | 18:28 |
| jeblair | dprince: sure, make a patch to pip-requires in nova setting the version of python-novaclient it requires. | 18:28 |
| *** danishman has joined #openstack-dev | 18:28 | |
| jaypipes | winston-d: no probs | 18:28 |
| jeblair | dprince: i'm not sure the version of python-novaclient on pypi ever existed in git. | 18:29 |
| dprince | jeblair: There are only 2 version of python-novaclient in Pypi. | 18:29 |
| *** asdfasdf has joined #openstack-dev | 18:29 | |
| dprince | http://pypi.python.org/pypi/python-novaclient | 18:29 |
| dprince | That is where it is pulling it from. | 18:29 |
| dprince | jeblair: bzr maybe then? | 18:30 |
| jeblair | actually, it's getting a version called 2012.1 that glance specifies from launchpad... | 18:30 |
| *** dachary has quit IRC | 18:30 | |
| jeblair | it really is kind of a mess. | 18:30 |
| dprince | jeblair: Yes. I know. | 18:31 |
| * mtaylor injets into conversation | 18:31 | |
| dprince | jeblair: Diablo is looking specifically for python-novaclient==2.6.0 | 18:31 |
| dprince | And there is a file in that tarball from PyPi that is patched incorrectly. | 18:31 |
| dprince | Fix that patch... post a 2.6.1 and we may be Okay here. | 18:31 |
| *** nati has quit IRC | 18:32 | |
| mtaylor | dprince: alternately, can we just have diablo use current python-novaclient? | 18:32 |
| dprince | mtaylor: That won't work. | 18:32 |
| dprince | mtaylor: API's have changed. | 18:32 |
| mtaylor | dprince: epic fail | 18:32 |
| mtaylor | ok | 18:32 |
| dprince | mtaylor: Diablo is old school. Uses image ID's, flavor ID's. All the old time goodies. | 18:32 |
| jeblair | dprince: yes, sorry, you're right about requiring 2.6.0 specifically. then yes, i agree the right solution is to fix 2.6.0 or release 2.6.0.1 on pypi. | 18:33 |
| mtaylor | I think 2.6.0.1 | 18:33 |
| mtaylor | dprince: can you make the 2.6.0.1 patch? | 18:34 |
| dprince | Okay. The offending file is novaclient/v1_1/servers.py in that tarball. | 18:34 |
| dprince | mtaylor: Sure. | 18:34 |
| dprince | mtaylor: I suppose I wanted to go at it from the sources... and have this refer to a proper Git tag. | 18:34 |
| mtaylor | dprince: awesome. if you get me a thing that will work, I'll upload that as 2.6.0.1 - and then we'll all sit down and make a real client library process next week | 18:34 |
| mtaylor | dprince: yes, I think that's sensible ... we don't really have a stable/diablo branch of novaclient (since we weren't doing multi client lib versions) | 18:35 |
| russellb | so i downloaded 2.6.0 ... i don't know how it ever worked. it has merge conflicts in it | 18:35 |
| dprince | russellb: yes. The merge conflicts in novaclient/v1_1/servers.py are the problem. | 18:35 |
| mtaylor | dprince: but perhaps we should make one given the breakage involved here | 18:35 |
| vishy | tr3buchet: ok so you are ok with the code as is? | 18:36 |
| mtaylor | jeblair: ^^ ? | 18:36 |
| russellb | dprince: ah, k :-) | 18:36 |
| tr3buchet | vishy: this merge prop? | 18:36 |
| dprince | mtaylor: I'd say lets fix this and then push it to a branch. For reference. | 18:36 |
| dprince | mtaylor: It sounds like Diablo will hang around until we cut Folsom. | 18:37 |
| mtaylor | dprince: ok | 18:37 |
| jeblair | well, problem is, we did not intend to have multiple branches of client libraries. | 18:37 |
| jeblair | because the client libraries are always supposed to work with all versions of the software | 18:37 |
| jeblair | except, apparently, diablo | 18:37 |
| dprince | That just isn't the case. | 18:38 |
| jeblair | which never really existed in the eyes of some people. | 18:38 |
| dprince | Well. The Nova v1.1 API was a moving target across the Diablo Essex releases. So that is what causes the pain here. | 18:38 |
| jeblair | yep | 18:38 |
| jeblair | at any rate, creating a 'stable/diablo' branch of novaclient may have unintented consequences, since several other processes in play weren't designed for it. | 18:40 |
| jeblair | we could create it, and see if things work | 18:40 |
| vishy | tr3buchet: yes the one for updating the cache in two places | 18:41 |
| vishy | tr3buchet: also can you review this one: https://review.openstack.org/#change,6396,patchset=3 | 18:41 |
| *** bepernoot has joined #openstack-dev | 18:41 | |
| jeblair | or we could create another branch, push the 2.6.0.1 commit to it, tag it, delete it, release to pypi and call it done until the summit. | 18:41 |
| *** berendt has quit IRC | 18:41 | |
| dprince | jeblair: I'm fine w/ that sir. Let me have a look at a patch... | 18:42 |
| *** bepernoot has quit IRC | 18:43 | |
| jeblair | mtaylor: ^ sound good? | 18:43 |
| jeblair | dprince: what's the sha1 you want to base your patch on? | 18:44 |
| dprince | jeblair: I get to decide! | 18:45 |
| dprince | jeblair: I suppose I was just going to download the tarball and diff the files... | 18:45 |
| mtaylor | jeblair: yes | 18:45 |
| dprince | jeblair: I have no idea what Git hash 2.6.0 came from. | 18:46 |
| jeblair | okay, so two things we can do: | 18:46 |
| *** pixelbeat has quit IRC | 18:46 | |
| dprince | jeblair: I'm going to look in LP too... if I can find it. | 18:46 |
| dprince | the hash that is... | 18:47 |
| jeblair | 1) if we want to get this into the git history, pick the closest sha1 to 2.6.0, i'll make a branch, and you can submit a change to gerrit for it; then we'll cut a release, tag that branch 2.6.0.1 and delete it. | 18:47 |
| jeblair | 2) we can just go along with the mysterious 2.6.0, patch it, email that to monty and he can upload 2.6.0.1 to pypi. | 18:47 |
| dprince | https://launchpad.net/ubuntu/oneiric/+source/python-novaclient | 18:47 |
| sdague | I'm trying to test the nova migration code on mysql, but the mysql driver isn't put in the venv it seems. What's the correct way to get that added so that this will run? | 18:48 |
| *** davlaps has quit IRC | 18:48 | |
| jeblair | sdague: add it to pip-requires if it's needed to run, or test-requires if it's only needed for testing, but, um, mysql seems like it's a pretty basic dependency and should already be there..? | 18:50 |
| mtaylor | jeblair: we do sqlite in the normal test runs | 18:51 |
| sdague | jeblair: yeh, it's not | 18:51 |
| sdague | mtaylor: on the jenkins server do you enable the mysql driver? | 18:51 |
| mtaylor | sdague: pip install python-mysqldb into your venv | 18:51 |
| mtaylor | sdague: no | 18:51 |
| mtaylor | sdague: we just run the test suite as is in the tree | 18:52 |
| sdague | mtaylor: can we get that changed? I was about to put in a unit test to ensure the mysql tables are all innodb, but if it never gets run, it's not very effective | 18:52 |
| jeblair | mtaylor: ah, gotcha. | 18:52 |
| mtaylor | sdague: nope | 18:52 |
| mtaylor | sdague: we use sqlite in the unittests because using mysql involves a bunch of setup/auth stuff | 18:53 |
| jeblair | sdague: then nevermind about the -requires files. | 18:53 |
| *** kindaopsdevy_ has joined #openstack-dev | 18:53 | |
| jeblair | dprince, mtaylor: going to grab lunch; let me know if you want to make that branch, or just email monty a patch if you want to go the other way. | 18:54 |
| *** ncode has quit IRC | 18:54 | |
| sdague | mtaylor: the way I solved that in another project was to have some default mysql testing user / db, and just skip test if that didn't match on the environment | 18:54 |
| mtaylor | sdague: for the most part, mysql gets tested via the devstack runs (which do buld on top of mysql) | 18:55 |
| sdague | mtaylor: there are enough parts of the migration that are mysql specific, and mysql is the database people are running with, I'd be really wary of not having that part of standard testing | 18:55 |
| sdague | mtaylor: right, but that won't catch and stop a bad migration until it's already in the tree | 18:56 |
| mtaylor | sdague: thats not true | 18:56 |
| mtaylor | sdague: we run devstack integration testing before every commit lands | 18:56 |
| mtaylor | sdague: however, I think you're pointing out that we need to test upgrades | 18:56 |
| mtaylor | sdague: also, don't get me wrong - I agree with your sentiment :) | 18:56 |
| sdague | mtaylor: :) | 18:56 |
| mtaylor | sdague: we need to test postgres as well, which is getting no love at the moment | 18:56 |
| sdague | mtaylor: well, what are your thoughts on the skip test approach? | 18:57 |
| jeblair | hey, this reminded me we have an openstack-ci meeting starting in 3 minutes | 18:57 |
| sdague | should I just take this over ther? | 18:57 |
| mtaylor | sdague: I'm not opposed in principle to the skip test approach - although making sure that we don't accidentally bork someone's db would be important | 18:57 |
| mtaylor | yeah | 18:57 |
| sdague | ok, cool, lets just continue in that meeting. | 18:58 |
| *** Ahmad_HP has quit IRC | 18:59 | |
| dprince | mtaylor/jeblair: I sent Monty the email w/ a patch. | 19:01 |
| *** hashar has joined #openstack-dev | 19:01 | |
| mtaylor | dprince: thanks! | 19:03 |
| *** bsza has joined #openstack-dev | 19:06 | |
| *** sandywalsh has quit IRC | 19:07 | |
| *** vincentricci has quit IRC | 19:08 | |
| zul | mtaylor: i know there are some users who really like to use postgresql rather than mysql | 19:14 |
| tr3buchet | vishy: 6396 is good | 19:15 |
| tr3buchet | vishy: i don't think i understand why it has to be updated from two places.. | 19:16 |
| tr3buchet | vishy: why not just update it for both associate and disassociate in the compute api | 19:16 |
| vishy | tr3buchet: um… that is what i was asking you about from the beginning | 19:18 |
| *** andrewsben is now known as andrewsben_zz | 19:19 | |
| vishy | tr3buchet: disassociate doesn't go through the compute api | 19:19 |
| tr3buchet | vishy: ah! makes sense. they are both called through the network api... seems like it should be done there... | 19:19 |
| *** dachary has joined #openstack-dev | 19:19 | |
| vishy | tr3buchet: well i guess that is ok. Officially it should be computes job since it is computes cache | 19:20 |
| vishy | but if we want to put both updates there for consistency/small set of changes that seems fine | 19:20 |
| *** sandywalsh has joined #openstack-dev | 19:20 | |
| vishy | tr3buchet: commented suggesting we do it in network_api | 19:23 |
| vishy | tr3buchet: thanks | 19:23 |
| *** andrewsben_zz is now known as andrewsben | 19:23 | |
| tr3buchet | vishy: i see what you mean... | 19:24 |
| tr3buchet | vishy: that gets screwy multi-cell | 19:24 |
| tr3buchet | vishy: this may need to be revisited later, but for now, the network api will update the cache in whatever cell the caller is in | 19:27 |
| *** dprince has quit IRC | 19:28 | |
| *** andrewsben is now known as andrewsben_zz | 19:28 | |
| *** armaan has left #openstack-dev | 19:30 | |
| *** bepernoot has joined #openstack-dev | 19:31 | |
| *** bepernoot has quit IRC | 19:35 | |
| mtaylor | zul: yes - but postgres is not getting any love in the gating | 19:36 |
| zul | mtaylor: i would love to see that changed | 19:36 |
| *** asdfasdf has quit IRC | 19:37 | |
| *** andrewsben_zz is now known as andrewsben | 19:37 | |
| mtaylor | zul: I'd also love to see maas written in such a way that it could run on multiple database backends :) | 19:39 |
| mtaylor | zul: you know, using something standard, like sqlalchemy? :) | 19:40 |
| zul | mtaylor: it can apparently | 19:40 |
| mtaylor | zul: great. then i'd like to see the packaging fixed so that it doens't install postgres on my laptop when I try to check i tout | 19:40 |
| zul | mtaylor: submit a bug then :) | 19:40 |
| mtaylor | zul: should I file a bug? :) | 19:40 |
| zul | mtaylor: yes | 19:40 |
| *** tryggvil has quit IRC | 19:40 | |
| mtaylor | zul: also - fwiw, it doesn't work on laptops running network-manager very well | 19:41 |
| zul | mtaylor: submit a bug please :) | 19:41 |
| mtaylor | zul: and current network-manager on desktop has taken over dnsmasq in such a way that using one desktop laptop to netboot another one which does not have a cd drive is now way harder | 19:41 |
| mtaylor | zul: I will | 19:41 |
| mtaylor | zul: but it's more fun to just tell you :) | 19:41 |
| adam_g | is glance backed by swift known to be working? seems i can't get it to function properly without doing some hacking to configure and use a tenant associated with swift_store_user when setting up its swift client | 19:42 |
| adam_g | jaypipes: ^ | 19:42 |
| *** danwent has joined #openstack-dev | 19:46 | |
| jaypipes | adam_g: yes, Glance + Swift has long been working... | 19:47 |
| jaypipes | adam_g: pastebin your conf... | 19:47 |
| jaypipes | adam_g: redact your sensitive parts :) | 19:47 |
| openstackgerrit | Verification of a change to openstack/nova failed: Fix errors in os-networks extension https://review.openstack.org/6396 | 19:49 |
| adam_g | jaypipes: ive just basically used devstack, set default_store = swift, and populated the credentials with those set by devstack + set it to create the container. i get this when trying to do a glance add: http://paste.ubuntu.com/923870/ | 19:49 |
| adam_g | jaypipes: quickly adding a tenant config option and passing it to swift client during instantiation seems to get everything working | 19:50 |
| *** bepernoot has joined #openstack-dev | 19:51 | |
| jaypipes | adam_g: I think this may be a bug in devstack then... mind submitting the above for devstack? | 19:53 |
| jaypipes | adam_g: just mention in the bug report about tenant needing to be set by devstack's writing the glance-api.conf | 19:53 |
| adam_g | jaypipes: how is it set in glance-api.conf? i dont see an option, and had to add one myself? | 19:55 |
| adam_g | and also didn't see it being passed to swift client at all | 19:56 |
| *** eglynn has joined #openstack-dev | 19:56 | |
| *** misheska has joined #openstack-dev | 19:57 | |
| openstackgerrit | Verification of a change to openstack/python-novaclient failed: Fixed the subcommand error message for nova shell. https://review.openstack.org/5819 | 19:59 |
| *** Ryan_Lane has quit IRC | 20:07 | |
| *** Ryan_Lane has joined #openstack-dev | 20:08 | |
| dtroyer | adam_g: could you check out https://review.openstack.org/5988 against a current devstack and let me know if it solves your problem? | 20:08 |
| *** maplebed is now known as maplebedAFK | 20:09 | |
| *** rkukura has joined #openstack-dev | 20:09 | |
| dtroyer | adam_g: hmmm…looking closer it may need to be rebased since we changed how glance is configured. but the idea is sound | 20:10 |
| adam_g | dtroyer: i can try, but im curious to know if swift store is expected to work with no use of tenant | 20:10 |
| adam_g | dtroyer: ah | 20:11 |
| adam_g | dtroyer: nvm, i see the delta in how ive configured it. one sec | 20:11 |
| adam_g | (more interested in seeing this work without devstack) | 20:11 |
| adam_g | dtroyer: still, seems the tenant is not parsed and correctly passed to swift client | 20:15 |
| zykes- | anyopne here from dell ? | 20:18 |
| *** rkukura has quit IRC | 20:18 | |
| *** bepernoot has quit IRC | 20:18 | |
| *** novas0x2a|laptop has quit IRC | 20:19 | |
| *** Ryan_Lane has quit IRC | 20:21 | |
| *** Ryan_Lane has joined #openstack-dev | 20:21 | |
| *** bepernoot has joined #openstack-dev | 20:22 | |
| *** utlemming has quit IRC | 20:23 | |
| *** utlemming has joined #openstack-dev | 20:27 | |
| *** danishman has quit IRC | 20:31 | |
| *** markwash has joined #openstack-dev | 20:31 | |
| *** paulormg has quit IRC | 20:33 | |
| *** andrewsben is now known as andrewsben_zz | 20:33 | |
| *** markwash has joined #openstack-dev | 20:34 | |
| vishy | comstud: https://review.openstack.org/#change,6396 | 20:34 |
| *** novas0x2a|laptop has joined #openstack-dev | 20:34 | |
| comstud | lookin | 20:35 |
| comstud | vishy: ahhh | 20:38 |
| annegentle | anyone know how do you get the aggregate_id needed for the os_aggregates extension? | 20:40 |
| openstackgerrit | Verification of a change to openstack/glance failed: Skip test depending on sqlite3 if unavailable https://review.openstack.org/6261 | 20:40 |
| rmk | So, boot from volume works, so long as I populate the volume with a bootable OS first | 20:41 |
| rmk | I'm guessing that's by design, in that there's no integration yet with glance | 20:41 |
| *** dwalleck has quit IRC | 20:43 | |
| *** bepernoot has quit IRC | 20:45 | |
| *** jakedahn_zz is now known as jakedahn | 20:45 | |
| *** joesavak has joined #openstack-dev | 20:45 | |
| chmouel | adam_g: what swift version are you using ? | 20:46 |
| *** andrewsben_zz is now known as andrewsben | 20:50 | |
| mikal | vishy: any chance I could get you to take a look at https://review.openstack.org/#change,6354 sometime? | 20:56 |
| *** gabrielhurley has joined #openstack-dev | 21:00 | |
| *** hashar has quit IRC | 21:00 | |
| *** dolphm has joined #openstack-dev | 21:00 | |
| *** andrewsben is now known as andrewsben_zz | 21:06 | |
| *** vincentricci has joined #openstack-dev | 21:08 | |
| *** rkukura has joined #openstack-dev | 21:10 | |
| adam_g | chmouel: trunk of master | 21:10 |
| chmouel | adam_g: what do you have in your glance configuration for swift? | 21:10 |
| adam_g | chmouel: http://paste.ubuntu.com/923983/ | 21:13 |
| adam_g | comstud: ah, s/35357/5000/ on that config snippet. traceback here http://paste.ubuntu.com/923985/ | 21:15 |
| adam_g | chmouel: which is supposed to be responsible for parsing tenant from user:tenant, swift or glance? | 21:16 |
| chmouel | swift client.py | 21:17 |
| *** maplebedAFK is now known as maplebed | 21:18 | |
| chmouel | adam_g: I'm checking your error you are getting | 21:18 |
| chmouel | adam_g: oh my review was integrated | 21:19 |
| chmouel | adam_g: i'll fix glance | 21:20 |
| *** Ryan_Lane has quit IRC | 21:20 | |
| *** Ryan_Lane1 has joined #openstack-dev | 21:20 | |
| chmouel | jaypipes, bcwaldon: this doesn't affect the release only trunk | 21:21 |
| adam_g | chmouel: huh? | 21:21 |
| *** gabrielhurley has quit IRC | 21:21 | |
| *** gabrielhurley1 has joined #openstack-dev | 21:21 | |
| adam_g | chmouel: so using essex/stable of swift + glance is expected to work? | 21:22 |
| chmouel | adam_g: this patch https://github.com/openstack/swift/commit/208b8e85a80e46ddb49dc2035cb292570a20c7db was merged | 21:22 |
| chmouel | adam_g: yes | 21:22 |
| *** joesavak has quit IRC | 21:22 | |
| adam_g | ah, i see | 21:22 |
| adam_g | thanks chmouel | 21:22 |
| chmouel | adam_g: just point your stackrc to stable | 21:22 |
| markmc | vishy, hmm, looks like stable-maint can't approve essex nominations for nova | 21:22 |
| markmc | vishy, intentional? we can for glance | 21:22 |
| chmouel | adam_g: and get that other devstack review for glance and swift+keystone support | 21:23 |
| *** gabrielhurley has joined #openstack-dev | 21:23 | |
| adam_g | yup | 21:23 |
| chmouel | adam_g: cool! ping me if that's broken | 21:23 |
| vishy | markmc: probably not intentional | 21:23 |
| jaypipes | adam_g: sorry, went to pay my taxes... :( | 21:25 |
| *** pixelbeat has joined #openstack-dev | 21:25 | |
| *** gabrielhurley1 has quit IRC | 21:25 | |
| adam_g | jaypipes: np | 21:26 |
| *** Ryan_Lane1 has quit IRC | 21:26 | |
| *** Ryan_Lane has joined #openstack-dev | 21:26 | |
| *** utlemming has quit IRC | 21:26 | |
| dtroyer | mtaylor: are you the guys who can make a stable/essex branch for devstack? | 21:27 |
| *** kindaopsdevy_ has quit IRC | 21:28 | |
| *** dwalleck has joined #openstack-dev | 21:30 | |
| *** utlemming has joined #openstack-dev | 21:31 | |
| adam_g | speaking of which, is there going to be a stable/essex branch of tempest? | 21:33 |
| *** rkukura has quit IRC | 21:34 | |
| *** rkukura has joined #openstack-dev | 21:34 | |
| *** rkukura has quit IRC | 21:35 | |
| *** rkukura has joined #openstack-dev | 21:39 | |
| *** jakedahn is now known as jakedahn_zz | 21:40 | |
| jaypipes | dtroyer: yes, or jeblair | 21:43 |
| *** PaulM has joined #openstack-dev | 21:45 | |
| *** lorin1 has quit IRC | 21:46 | |
| mtaylor | dtroyer: I am - but jeblair is smarter than I am | 21:47 |
| mtaylor | dtroyer: btw - I'm excited about your one-client thoughts | 21:47 |
| dtroyer | mtaylor: thx; I think it'll be interesting next week… | 21:51 |
| mtaylor | dtroyer: I almost hacked together a proof of concept this weekend, but then realized I had better things to do with my time | 21:52 |
| *** markmc has quit IRC | 21:52 | |
| *** slyphon has quit IRC | 22:00 | |
| *** paulstallworth has quit IRC | 22:02 | |
| *** nati has joined #openstack-dev | 22:02 | |
| *** jakedahn_zz is now known as jakedahn | 22:03 | |
| *** jakedahn is now known as jakedahn_zz | 22:05 | |
| *** dolphm has quit IRC | 22:06 | |
| *** lloydde has quit IRC | 22:07 | |
| *** rkukura has quit IRC | 22:07 | |
| *** novas0x2a|lapto1 has joined #openstack-dev | 22:08 | |
| *** novas0x2a|lapto1 has quit IRC | 22:08 | |
| *** novas0x2a|lapto1 has joined #openstack-dev | 22:09 | |
| *** novas0x2a|laptop has quit IRC | 22:09 | |
| *** rkukura has joined #openstack-dev | 22:10 | |
| *** novas0x2a|lapto1 has quit IRC | 22:10 | |
| *** novas0x2a|laptop has joined #openstack-dev | 22:14 | |
| *** crobinso has quit IRC | 22:14 | |
| *** bsza has quit IRC | 22:14 | |
| *** asdfasdf has joined #openstack-dev | 22:20 | |
| *** nati has quit IRC | 22:21 | |
| jeblair | dtroyer: yes, can do stable/essex for devstack. what sha1 would you like for that? | 22:27 |
| jeblair | (curren master?) | 22:28 |
| jeblair | current, even | 22:28 |
| dtroyer | jeblair: yes, current. | 22:28 |
| jeblair | dtroyer: done | 22:29 |
| *** kbringard has quit IRC | 22:29 | |
| dtroyer | jeblair: I'll have one commit to add to it to change the branches in stackrc…how do I target that? | 22:29 |
| jeblair | 'git review stable/essex' | 22:29 |
| dtroyer | cool, thanks | 22:29 |
| jeblair | that answer your question? | 22:29 |
| dtroyer | yup | 22:29 |
| *** asdfasdf has quit IRC | 22:31 | |
| *** rkukura has quit IRC | 22:31 | |
| *** aswin has joined #openstack-dev | 22:34 | |
| *** aswin has left #openstack-dev | 22:34 | |
| *** dwalleck has quit IRC | 22:34 | |
| *** Ryan_Lane1 has joined #openstack-dev | 22:37 | |
| *** Ryan_Lane has quit IRC | 22:37 | |
| *** Ryan_Lane1 has quit IRC | 22:37 | |
| *** Ryan_Lane1 has joined #openstack-dev | 22:37 | |
| *** Ryan_Lane1 is now known as Ryan_Lane | 22:37 | |
| *** Shrews has quit IRC | 22:46 | |
| *** nati has joined #openstack-dev | 22:47 | |
| *** nati has quit IRC | 22:47 | |
| *** Shrews has joined #openstack-dev | 22:48 | |
| *** mattray has quit IRC | 22:48 | |
| *** pixelbeat has quit IRC | 22:50 | |
| *** novas0x2a|laptop has quit IRC | 22:50 | |
| *** novas0x2a|laptop has joined #openstack-dev | 22:50 | |
| *** dtroyer is now known as dtroyer_zzz | 22:55 | |
| *** GheRivero has quit IRC | 23:05 | |
| *** rnirmal has quit IRC | 23:08 | |
| *** Ryan_Lane has quit IRC | 23:12 | |
| *** Ryan_Lane has joined #openstack-dev | 23:12 | |
| *** roge has quit IRC | 23:21 | |
| *** dtroyer_zzz is now known as dtroyer | 23:27 | |
| *** Ryan_Lane has quit IRC | 23:39 | |
| *** Ryan_Lane has joined #openstack-dev | 23:39 | |
| *** heckj has quit IRC | 23:40 | |
| *** paulstallworth has joined #openstack-dev | 23:44 | |
| *** slyphon has joined #openstack-dev | 23:47 | |
| *** lloydde has joined #openstack-dev | 23:48 | |
| *** winston-d has quit IRC | 23:49 | |
| *** dachary has quit IRC | 23:52 | |
| *** Ryan_Lane has quit IRC | 23:58 | |
| *** Ryan_Lane has joined #openstack-dev | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!