| *** dachary has joined #openstack-dev | 00:00 | |
| *** e1mer has joined #openstack-dev | 00:04 | |
| *** galstrom has joined #openstack-dev | 00:07 | |
| *** zhuadl has joined #openstack-dev | 00:10 | |
| *** dachary has quit IRC | 00:17 | |
| *** sc68cal has joined #openstack-dev | 00:18 | |
| *** sc68cal has joined #openstack-dev | 00:18 | |
| *** dtroyer_zzz is now known as dtroyer | 00:29 | |
| *** harlowja has left #openstack-dev | 00:29 | |
| *** dachary has joined #openstack-dev | 00:30 | |
| *** Kiall has quit IRC | 00:31 | |
| *** s0mik has joined #openstack-dev | 00:31 | |
| *** dtroyer is now known as dtroyer_zzz | 00:33 | |
| *** halfss has quit IRC | 00:34 | |
| *** Kiall has joined #openstack-dev | 00:35 | |
| *** mattray has joined #openstack-dev | 00:35 | |
| *** dachary has quit IRC | 00:35 | |
| *** jtran has quit IRC | 00:36 | |
| *** ncode has quit IRC | 00:39 | |
| *** andrewbogott has quit IRC | 00:39 | |
| *** torgomatic has quit IRC | 00:42 | |
| *** galstrom has quit IRC | 00:42 | |
| *** anniec has quit IRC | 00:49 | |
| *** dachary has joined #openstack-dev | 00:49 | |
| *** mattray has quit IRC | 00:54 | |
| *** mattray has joined #openstack-dev | 00:55 | |
| *** dachary has quit IRC | 00:55 | |
| *** Gordonz has joined #openstack-dev | 00:57 | |
| *** heckj has quit IRC | 00:58 | |
| *** dachary has joined #openstack-dev | 00:58 | |
| mnaser | nova stop deletes the vm with xenserver? is there a way to make it just shutdown? | 00:58 |
|---|---|---|
| *** Aaton is now known as Aaton_off | 01:03 | |
| *** winston-d has joined #openstack-dev | 01:04 | |
| *** dachary has quit IRC | 01:04 | |
| *** dachary has joined #openstack-dev | 01:05 | |
| *** hugokuo1 has joined #openstack-dev | 01:08 | |
| *** mnaser has quit IRC | 01:10 | |
| *** hugokuo has quit IRC | 01:11 | |
| *** PotHix has quit IRC | 01:11 | |
| *** dachary has quit IRC | 01:11 | |
| *** Mandell has quit IRC | 01:13 | |
| *** zhuadl has quit IRC | 01:18 | |
| *** johnpostlethwait has quit IRC | 01:23 | |
| *** ayoung has quit IRC | 01:24 | |
| *** jakedahn is now known as jakedahn_zz | 01:25 | |
| *** dtroyer_zzz is now known as dtroyer | 01:27 | |
| *** jog0 has quit IRC | 01:27 | |
| *** dachary has joined #openstack-dev | 01:37 | |
| *** dachary has quit IRC | 01:43 | |
| *** ncode has joined #openstack-dev | 01:44 | |
| *** ncode has joined #openstack-dev | 01:44 | |
| *** johnpur has quit IRC | 01:45 | |
| *** hugokuo has joined #openstack-dev | 01:47 | |
| *** hugokuo1 has quit IRC | 01:50 | |
| *** zhuadl has joined #openstack-dev | 01:50 | |
| *** sacharya has joined #openstack-dev | 01:52 | |
| dryan_installing | ping pmyers | 01:55 |
| *** dryan_installing is now known as dyan | 01:55 | |
| *** galstrom has joined #openstack-dev | 01:56 | |
| *** dyan is now known as dryan | 01:56 | |
| *** novas0x2a|laptop has quit IRC | 02:05 | |
| *** s0mik has quit IRC | 02:08 | |
| *** dachary has joined #openstack-dev | 02:11 | |
| *** littleidea has joined #openstack-dev | 02:19 | |
| *** littleidea has quit IRC | 02:28 | |
| *** bencherian has quit IRC | 02:31 | |
| *** dachary has quit IRC | 02:32 | |
| *** otter1 has quit IRC | 02:38 | |
| *** Gordonz has quit IRC | 02:43 | |
| *** salgado-afk has quit IRC | 02:43 | |
| *** otter1 has joined #openstack-dev | 02:45 | |
| *** roge has quit IRC | 02:47 | |
| *** s0mik has joined #openstack-dev | 02:53 | |
| *** edygarcia has quit IRC | 02:55 | |
| *** pixelbeat has quit IRC | 02:57 | |
| *** hugokuo1 has joined #openstack-dev | 03:00 | |
| *** dachary has joined #openstack-dev | 03:02 | |
| *** hugokuo has quit IRC | 03:02 | |
| *** dachary has quit IRC | 03:07 | |
| *** galstrom has quit IRC | 03:15 | |
| *** ncode has quit IRC | 03:15 | |
| *** Mandell has joined #openstack-dev | 03:22 | |
| *** salv-orlando has quit IRC | 03:28 | |
| *** galstrom has joined #openstack-dev | 03:29 | |
| *** cmagina has quit IRC | 03:32 | |
| *** anniec has joined #openstack-dev | 03:33 | |
| *** cmagina has joined #openstack-dev | 03:34 | |
| *** salv-orlando has joined #openstack-dev | 03:46 | |
| *** mokas_ has quit IRC | 03:47 | |
| *** william0 has joined #openstack-dev | 03:49 | |
| *** salv-orlando_ has joined #openstack-dev | 03:52 | |
| *** salv-orlando has quit IRC | 03:52 | |
| *** salv-orlando_ is now known as salv-orlando | 03:52 | |
| *** salv-orlando_ has joined #openstack-dev | 03:54 | |
| *** salv-orlando has quit IRC | 03:54 | |
| *** salv-orlando_ is now known as salv-orlando | 03:54 | |
| *** markmcclain has quit IRC | 03:58 | |
| *** mnewby has joined #openstack-dev | 03:58 | |
| *** thingee has quit IRC | 04:00 | |
| *** salv-orlando has quit IRC | 04:02 | |
| *** mnewby_ has joined #openstack-dev | 04:02 | |
| *** mnewby has quit IRC | 04:04 | |
| *** mnewby_ is now known as mnewby | 04:04 | |
| *** zhuadl has quit IRC | 04:14 | |
| *** sacharya has quit IRC | 04:25 | |
| *** dubsquared has joined #openstack-dev | 04:33 | |
| *** mnewby has quit IRC | 04:33 | |
| *** steveb_ has quit IRC | 04:40 | |
| *** mokas has joined #openstack-dev | 04:46 | |
| *** galstrom has quit IRC | 04:53 | |
| *** dubsquared1 has joined #openstack-dev | 05:02 | |
| *** mokas has quit IRC | 05:05 | |
| *** dubsquared has quit IRC | 05:05 | |
| *** garyk has quit IRC | 05:05 | |
| *** william0 has quit IRC | 05:10 | |
| *** thingee has joined #openstack-dev | 05:12 | |
| *** thingee is now known as thingee_zzz | 05:12 | |
| *** dubsquared1 has quit IRC | 05:20 | |
| *** zaitcev has quit IRC | 05:38 | |
| *** mokas has joined #openstack-dev | 05:39 | |
| *** william0 has joined #openstack-dev | 05:47 | |
| *** dtroyer is now known as dtroyer_zzz | 05:47 | |
| *** william0 has quit IRC | 05:49 | |
| *** william0 has joined #openstack-dev | 05:50 | |
| *** zhuadl has joined #openstack-dev | 06:00 | |
| *** mrunge has joined #openstack-dev | 06:03 | |
| *** anniec has quit IRC | 06:06 | |
| *** GheRivero has joined #openstack-dev | 06:16 | |
| *** s0mik has quit IRC | 06:18 | |
| *** danwent has joined #openstack-dev | 06:19 | |
| *** garyk has joined #openstack-dev | 06:23 | |
| *** blamar_ has joined #openstack-dev | 06:25 | |
| *** blamar has quit IRC | 06:26 | |
| *** blamar_ is now known as blamar | 06:26 | |
| ttx | early reminder: today is bug squashing day, join us on #openstack-bugday if you participate to the effort: http://wiki.openstack.org/BugDays/20120712BugSquashing | 06:28 |
| garyk | danwent: ping | 06:30 |
| *** mrunge_ has joined #openstack-dev | 06:33 | |
| *** mrunge has quit IRC | 06:33 | |
| *** haleyb has quit IRC | 06:36 | |
| GheRivero | morning people | 06:38 |
| *** andrewsmedina has quit IRC | 06:39 | |
| *** mrunge_ is now known as mrunge | 06:40 | |
| *** mrunge has quit IRC | 06:40 | |
| *** mrunge has joined #openstack-dev | 06:41 | |
| *** haleyb has joined #openstack-dev | 06:45 | |
| *** rafaduran has quit IRC | 06:52 | |
| *** rafaduran has joined #openstack-dev | 06:52 | |
| *** reidrac has joined #openstack-dev | 07:10 | |
| *** eglynn_ has joined #openstack-dev | 07:10 | |
| *** erikzaadi has joined #openstack-dev | 07:15 | |
| *** dachary has joined #openstack-dev | 07:18 | |
| *** EmilienM has joined #openstack-dev | 07:19 | |
| *** m4xmr has joined #openstack-dev | 07:19 | |
| *** m4xmr has left #openstack-dev | 07:36 | |
| *** smoser has quit IRC | 07:37 | |
| hugokuo1 | Hi all , | 07:41 |
| hugokuo1 | does anyone play with swift | 07:42 |
| hugokuo1 | I wondering to know that relation between XFS and arp_cache | 07:42 |
| hugokuo1 | I saw that million of OBJECTS in slab"s arp_cahe | 07:42 |
| hugokuo1 | how to flush the arp_cache usage ? | 07:43 |
| *** mindpixel has joined #openstack-dev | 07:45 | |
| *** smoser has joined #openstack-dev | 07:46 | |
| *** dryan has quit IRC | 07:47 | |
| *** winston-d has quit IRC | 07:48 | |
| zykes- | EmilienM: ping | 07:51 |
| *** darraghb has joined #openstack-dev | 08:00 | |
| *** steveb_ has joined #openstack-dev | 08:14 | |
| *** derekh has joined #openstack-dev | 08:18 | |
| *** tserong has quit IRC | 08:21 | |
| *** dachary has quit IRC | 08:24 | |
| *** tserong has joined #openstack-dev | 08:24 | |
| *** tserong has joined #openstack-dev | 08:24 | |
| *** cmagina has quit IRC | 08:32 | |
| *** cmagina has joined #openstack-dev | 08:34 | |
| *** erikzaadi has quit IRC | 08:58 | |
| *** dachary has joined #openstack-dev | 09:06 | |
| *** danpb has joined #openstack-dev | 09:12 | |
| *** rods has joined #openstack-dev | 09:14 | |
| *** steveb_ has quit IRC | 09:16 | |
| *** steveb_ has joined #openstack-dev | 09:17 | |
| *** william0 has quit IRC | 09:24 | |
| *** m4xmr_ has joined #openstack-dev | 09:36 | |
| *** m4xmr_ is now known as m4xmr | 09:36 | |
| *** steveb_ has quit IRC | 09:38 | |
| *** steveb_ has joined #openstack-dev | 09:39 | |
| *** dryan has joined #openstack-dev | 09:41 | |
| danpb | jeblair: ping (about that devstack failure with libvirt) | 09:43 |
| *** amotoki has joined #openstack-dev | 09:49 | |
| *** amotoki has joined #openstack-dev | 09:50 | |
| *** steveb_ has quit IRC | 09:58 | |
| *** davidha has joined #openstack-dev | 10:04 | |
| *** dryan has quit IRC | 10:07 | |
| *** pixelbeat has joined #openstack-dev | 10:08 | |
| *** zhuadl has quit IRC | 10:09 | |
| *** zhuadl has joined #openstack-dev | 10:17 | |
| *** Leander has joined #openstack-dev | 10:27 | |
| Leander | hello all, can someone tell me what's happening here: https://jenkins.openstack.org/job/gate-tempest-devstack-vm/3306/consoleFull | 10:28 |
| *** zhuadl has quit IRC | 10:43 | |
| *** Kiall has quit IRC | 10:45 | |
| zykes- | EmilienM: did your vm's get a ip when using floodlight and quantum ? | 10:48 |
| *** mjfork has joined #openstack-dev | 10:49 | |
| *** dayou has joined #openstack-dev | 10:50 | |
| *** Kiall has joined #openstack-dev | 10:51 | |
| *** Leander has quit IRC | 10:52 | |
| *** dachary has quit IRC | 10:59 | |
| *** lts has joined #openstack-dev | 11:22 | |
| *** dryan has joined #openstack-dev | 11:28 | |
| *** hugokuo2 has joined #openstack-dev | 11:42 | |
| *** sdague_ has joined #openstack-dev | 11:42 | |
| *** uvirtbot` has joined #openstack-dev | 11:43 | |
| *** LinuxJedi_ has joined #openstack-dev | 11:44 | |
| *** dryan has quit IRC | 11:44 | |
| *** rafaduran_ has joined #openstack-dev | 11:44 | |
| *** hugokuo1 has quit IRC | 11:45 | |
| *** sdague has quit IRC | 11:45 | |
| *** uvirtbot has quit IRC | 11:45 | |
| *** LinuxJedi has quit IRC | 11:45 | |
| *** rafaduran has quit IRC | 11:45 | |
| *** markvoelker has joined #openstack-dev | 11:48 | |
| *** galstrom has joined #openstack-dev | 12:00 | |
| *** mrunge has quit IRC | 12:10 | |
| *** salgado-afk has joined #openstack-dev | 12:13 | |
| *** salgado-afk has quit IRC | 12:13 | |
| *** salgado-afk has joined #openstack-dev | 12:13 | |
| *** salgado-afk has quit IRC | 12:15 | |
| *** salgado has joined #openstack-dev | 12:15 | |
| *** salgado has joined #openstack-dev | 12:15 | |
| *** erikzaadi has joined #openstack-dev | 12:15 | |
| *** alaski has joined #openstack-dev | 12:21 | |
| *** Ruetobas has quit IRC | 12:22 | |
| *** sandywalsh has joined #openstack-dev | 12:24 | |
| *** mestery_ is now known as mestery | 12:26 | |
| *** Ruetobas has joined #openstack-dev | 12:27 | |
| soren | For those who care, I just posted for review three patches for stable/essex. With those patches, tempest master passes against it. Whee! | 12:27 |
| *** rkukura has left #openstack-dev | 12:33 | |
| *** galstrom has quit IRC | 12:34 | |
| *** sacharya has joined #openstack-dev | 12:45 | |
| *** dhellmann_ has joined #openstack-dev | 12:45 | |
| *** dhellmann_ has quit IRC | 12:46 | |
| *** dhellmann has quit IRC | 12:49 | |
| annegentle | soren: now that's a good start to bug squash day | 12:49 |
| *** LinuxJedi_ is now known as LinuxJedi | 12:50 | |
| *** roge has joined #openstack-dev | 12:52 | |
| *** GheRivero has quit IRC | 12:52 | |
| *** dayou has quit IRC | 12:52 | |
| *** anniec has joined #openstack-dev | 12:55 | |
| *** dprince has joined #openstack-dev | 12:55 | |
| *** e1mer has quit IRC | 12:56 | |
| *** sacharya has quit IRC | 12:58 | |
| *** mindpixel has quit IRC | 12:58 | |
| *** erikzaadi has quit IRC | 13:01 | |
| *** mnewby has joined #openstack-dev | 13:01 | |
| *** RoboticCheese has quit IRC | 13:03 | |
| *** RoboticCheese has joined #openstack-dev | 13:04 | |
| *** m4xmr has quit IRC | 13:05 | |
| ttx | TZ reminder: today is bug squashing day, join us on #openstack-bugday if you participate to the effort: http://wiki.openstack.org/BugDays/20120712BugSquashing | 13:07 |
| *** ncode has joined #openstack-dev | 13:07 | |
| *** ncode has joined #openstack-dev | 13:07 | |
| *** anniec has quit IRC | 13:07 | |
| *** anniec has joined #openstack-dev | 13:07 | |
| *** m4xmr has joined #openstack-dev | 13:09 | |
| *** e1mer has joined #openstack-dev | 13:10 | |
| *** dprince_ has joined #openstack-dev | 13:12 | |
| *** dprince_ has quit IRC | 13:12 | |
| *** sdague_ is now known as sdague | 13:12 | |
| *** e1mer has quit IRC | 13:20 | |
| *** edygarcia has joined #openstack-dev | 13:21 | |
| *** edygarcia has quit IRC | 13:26 | |
| mtaylor | garyk: morning | 13:26 |
| *** ayoung has joined #openstack-dev | 13:28 | |
| *** japage has joined #openstack-dev | 13:29 | |
| *** e1mer has joined #openstack-dev | 13:30 | |
| *** cmagina has quit IRC | 13:33 | |
| *** cmagina has joined #openstack-dev | 13:33 | |
| *** dolphm has quit IRC | 13:33 | |
| *** sacharya has joined #openstack-dev | 13:34 | |
| *** mindpixel has joined #openstack-dev | 13:39 | |
| *** RoboticCheese has quit IRC | 13:39 | |
| *** RoboticCheese has joined #openstack-dev | 13:45 | |
| *** andrewbogott has joined #openstack-dev | 13:50 | |
| *** andrewbogott has joined #openstack-dev | 13:50 | |
| ttx | soren: could you comment on the right way to fix https://bugs.launchpad.net/nova/+bug/996840 ? | 13:52 |
| uvirtbot` | Launchpad bug 996840 in libvirt "Libvirt error when trying to mount ISCSI volumes" [Medium,Confirmed] | 13:52 |
| danpb | jeblair: russellb: ping (about that devstack / libvirt cpu model problem from last night ) | 13:55 |
| *** danwent has quit IRC | 13:56 | |
| *** zhuadl has joined #openstack-dev | 13:57 | |
| danpb | ttx: it is a combination of a QEMU and libvirt issue | 14:01 |
| ttx | danpb: right.. nothing to fix in Nova there, right ? | 14:01 |
| soren | ttx: Will do. | 14:02 |
| ttx | soren: if it's a combination of a QEMU and libvirt issue, just redirect the bug to the corresponding packages and invalidate the nova task | 14:02 |
| soren | danpb: Well, it works when usingn attach-disk in virsh. | 14:04 |
| soren | danpb: So clearly there's a way to make it work with the same versions of libvirt and qemu. | 14:05 |
| *** edygarcia has joined #openstack-dev | 14:05 | |
| soren | ttx: Don't close the nova bug. There's still work to do here. At the very least, we should rip out the virsh call as soon as we can. This bug would help us track that. | 14:06 |
| ttx | ok, just add a comment so that nobody closes it. It's bugsquashing day so the temptation is high | 14:06 |
| ttx | soren: ^ | 14:06 |
| eglynn_ | bcwaldon: quick question about the glance v2 API impl? | 14:07 |
| * soren adds another comment. | 14:07 | |
| danpb | soren: hmm, comment #2 shows that they're using human-monitor-command passthrough | 14:07 |
| danpb | soren: which contradicts what is shown in the logs in the initial description | 14:07 |
| eglynn_ | bcwaldon: I'm missing how (or if) the image status transitions from queued to active in the v2 case | 14:08 |
| soren | danpb: The proposed patch (which allegedly fixes it (I haven't verified this myself, though)) replaces the current code with a system() call that invokes virsh attach-disk. | 14:09 |
| *** markmcclain has joined #openstack-dev | 14:09 | |
| *** dubsquared has joined #openstack-dev | 14:11 | |
| soren | danpb: Assuming the Python bindings aren't broken in some way, we should be able to replicate whatever virsh does in attach-disk via the C API. | 14:11 |
| *** dolph_ has joined #openstack-dev | 14:11 | |
| soren | danpb: Right? virsh isn't privileged in any way, afaik. | 14:11 |
| *** dolph_ is now known as dolphm | 14:12 | |
| * soren enjoys how convenient it is to have danpb right here for these discussions :) | 14:12 | |
| garyk | LinuxJedi: ping | 14:19 |
| LinuxJedi | garyk: hi | 14:19 |
| *** dtroyer_zzz is now known as dtroyer | 14:20 | |
| garyk | LinuxJedi: hi, maybe you can help. Any idea why jenkins fails https://review.openstack.org/#/c/8650/ | 14:20 |
| *** GheRivero has joined #openstack-dev | 14:20 | |
| danpb | soren: yea that fix can't possibly be right | 14:20 |
| danpb | soren: whatever virsh can do, python bindings can do to | 14:21 |
| danpb | perhaps they're generating different XML | 14:21 |
| *** rbasak has quit IRC | 14:21 | |
| *** maoy has joined #openstack-dev | 14:21 | |
| danpb | their logs don't show what virsh attach-disk generated | 14:21 |
| * danpb notes that standing 100ft below a Eurofighter Typhoon when they turn on the afterburners really hurts your ears | 14:22 | |
| * danpb loves being in an office that can view the Farnborough airshow | 14:22 | |
| LinuxJedi | mtaylor: garyk's problem, is that an issue with stable/essex's setup.py? | 14:23 |
| garyk | LinuxJedi: thanks. i'll ping him | 14:23 |
| mtaylor | garyk, LinuxJedi: I just submitted a patch to quantum stable/essex that will fix it | 14:23 |
| mtaylor | garyk: land that patch, then yours (and others) should work again | 14:24 |
| *** zhuadl has quit IRC | 14:25 | |
| LinuxJedi | mtaylor: you are like a psychic ninja ;) | 14:25 |
| *** halfss has joined #openstack-dev | 14:27 | |
| garyk | mtaylor: thanks! | 14:27 |
| *** danwent has joined #openstack-dev | 14:27 | |
| garyk | mtaylor: jenkins is a bit grumpy today - he did not like what you did | 14:28 |
| mtaylor | garyk: /me goes to punch him in the face | 14:28 |
| garyk | mtaylor: at school we used to have a rugby war cry "punch him, kick him, ruck him in the head, we are the mighty....". maybe he just needs some tlc | 14:29 |
| mtaylor | garyk: tlc coming | 14:29 |
| mtaylor | garyk: :) | 14:29 |
| mtaylor | garyk: patch updated | 14:29 |
| garyk | mtaylor: thanks | 14:29 |
| mtaylor | garyk: also - you guys who went to the types of schools where one would play rugby have the most charming songs :) | 14:30 |
| *** cp16net is now known as cp16net|away | 14:30 | |
| ttx | soren: another one in your pile: https://bugs.launchpad.net/nova/+bug/871663 | 14:31 |
| soren | danpb: Well, the fix might actually do the trick. It's just that the code it replaces didnt do everything that attach-disk does. | 14:31 |
| uvirtbot` | Launchpad bug 871663 in nova "plugins/xenserver/networking/etc/xensource/scripts/ovs_configure_vif_flows.py has lots of hardcoded paths" [Undecided,Incomplete] | 14:31 |
| soren | ttx: /me looks | 14:31 |
| garyk | mtaylor: we were forced to play. :) and sing along when other played... | 14:31 |
| mtaylor | hehe | 14:31 |
| * mtaylor thinks we should replace IRC with a MUD server ... | 14:32 | |
| soren | ttx: Try as I might, I cannot bring myself to give a flying <beep> about that bug anymore. | 14:33 |
| *** andrewsmedina has joined #openstack-dev | 14:33 | |
| ttx | soren: heh, then close it ;) | 14:33 |
| *** rbasak has joined #openstack-dev | 14:34 | |
| ttx | soren: OP privilege | 14:34 |
| danpb | soren: even if it works, it is not suitable for merging into Nova IMHO - it must be rewritten to use the python apis not virsh - as its stands its just a hack because the person didn't want to properly investigate what was actally wrong | 14:34 |
| soren | danpb: Oh, absolutely. | 14:35 |
| *** heckj has joined #openstack-dev | 14:35 | |
| soren | danpb: My point is just that enough people have verified that it works that it seems plausible that it works, so if we replace the current code with a closer equivalent to what attach-disk in virsh does, we should be fine. attach-disk is, as you know, a virshism. It's a wrapper around.. err... attachDevice or something. | 14:38 |
| *** ncode has quit IRC | 14:38 | |
| * soren has to run | 14:38 | |
| ttx | soren: ok, will close that bug for you | 14:39 |
| ttx | ah, you did it | 14:39 |
| *** ncode has joined #openstack-dev | 14:40 | |
| dansmith | can someone approve? https://review.openstack.org/#/c/9547/ | 14:41 |
| jeblair | danpb: pong (and that's awesome, i was just watching the video of the 787 doing near-acrobatics at farnborough!) | 14:43 |
| garyk | mtaylor: saw that it failed again. sorry | 14:45 |
| danpb | soren: were you actually able to reproduce the failure seen yourself ? | 14:45 |
| *** galstrom has joined #openstack-dev | 14:46 | |
| *** rkukura has joined #openstack-dev | 14:47 | |
| *** lorin1 has joined #openstack-dev | 14:48 | |
| jeblair | danpb: do you have thoughts on https://bugs.launchpad.net/devstack/+bug/1023513 | 14:48 |
| uvirtbot` | Launchpad bug 1023513 in devstack "devstack exercises fail after nova cpu mode change " [Undecided,New] | 14:48 |
| jeblair | danpb: (do you want to fix it? :) | 14:48 |
| *** simo has joined #openstack-dev | 14:50 | |
| dolphm | heckj: easy review https://review.openstack.org/#/c/9716/ | 14:50 |
| trapni | please help, I have changed to name of my nova-network (no multi_host) hostname via --host=$HA_GW_NAME, restarted, but now nothing works anymore (even like spawning new VMs) because they all still use the old hostname - why? | 14:51 |
| danpb | jeblair: ok, yes, this is probably my ignorance, but I'm not understanding exactly what/where the error you're seeing is | 14:51 |
| maoy | danpb: we see occaionally long delay in calling libvirt listDomain calls ( 10s of seconds) with no apparent workload on Ubuntu 12.04. Do you have any insight? Thanks. | 14:51 |
| *** dhellmann has joined #openstack-dev | 14:51 | |
| danpb | jeblair: I mean i can see the xml is different, but where does the failure appear | 14:51 |
| jeblair | danpb: k, 1 sec. | 14:52 |
| *** Gordonz has joined #openstack-dev | 14:52 | |
| danpb | (i don't use devstack, so not too familiar with what it does) | 14:52 |
| bcwaldon | eglynn_: status is still on the list ;) | 14:53 |
| * heckj yawns and reviews | 14:53 | |
| *** datsun180b has joined #openstack-dev | 14:53 | |
| bcwaldon | heckj: do we BORE you | 14:54 |
| heckj | dolphm: really? Seems kinda arbitrary. That just hit your funny bone? | 14:54 |
| *** mattray has quit IRC | 14:54 | |
| heckj | bcwaldon: no, insomnia | 14:54 |
| jeblair | danpb: so the rough overview is that stack.sh configures all the components, and then exercise.sh runs some basic tests (which are in individual scripts, like euca.sh) | 14:54 |
| jeblair | danpb: so if you look at https://jenkins.openstack.org/job/gate-tempest-devstack-vm/3086/consoleFull | 14:54 |
| bcwaldon | heckj: convenient excuse | 14:54 |
| heckj | bcwaldon: every now and then, it slams me into the ground. Today is one of those days | 14:54 |
| *** Mandell has quit IRC | 14:54 | |
| jeblair | danpb: and skip to "17:41:16 Running euca" | 14:55 |
| * heckj yawns at bcwaldon | 14:55 | |
| bcwaldon | !! | 14:55 |
| openstack | bcwaldon: Error: "!" is not a valid command. | 14:55 |
| heckj | heh | 14:55 |
| *** dhellmann_ has joined #openstack-dev | 14:55 | |
| bcwaldon | you win this time | 14:55 |
| jeblair | danpb: you'll see the command line calls that attempt to spin up an instance, and eventually fail | 14:55 |
| heckj | the openstack bots will get you! | 14:55 |
| ayoung | termie, dolphm can one of you guys approve https://review.openstack.org/#/c/9240/ | 14:55 |
| bourke | all, does it take a while for gerrit to recognise new members of the openstack-cla group? Have been approved to join about 20 minutes ago but still getting 'A Contributer Agreement must be completed...' error when typing git review | 14:55 |
| jeblair | danpb: does that help? (also, out of curiosity, what do you use to test your changes, since you don't use devstack?) | 14:56 |
| *** dhellmann has quit IRC | 14:56 | |
| *** dhellmann_ is now known as dhellmann | 14:56 | |
| jeblair | bourke: it can take up to an hour. :( | 14:56 |
| danpb | jeblair: ah, so the ping errors are the "erroR" here | 14:56 |
| jeblair | danpb: yep | 14:57 |
| bourke | jeblair: no problem, as long as it's not something on my end :) | 14:57 |
| jeblair | danpb: basically, it just times out | 14:57 |
| danpb | jeblair: i don't suppose you have the corresponding nova compute.log file, since that's where the interesting info will be | 14:57 |
| *** Dr_Who has joined #openstack-dev | 14:57 | |
| jeblair | danpb: check out the list of files at https://jenkins.openstack.org/job/gate-tempest-devstack-vm/3086/? | 14:58 |
| jeblair | danpb: and see if 'screen-n-cpu.txt' or 'syslog.txt' has what you need | 14:58 |
| *** eglynn_ is now known as eglynn | 14:59 | |
| mtaylor | bcwaldon: hey! you wanna hear something fun? | 15:00 |
| danpb | jeblair: strange, neither of those show any signs of failure | 15:00 |
| bcwaldon | mtaylor: go for it | 15:00 |
| danpb | jeblair: i guess we should just make devstack set libvirt_cpu_mode="none" | 15:01 |
| mtaylor | bcwaldon: I figured out how to turn all of the openstack.common {setup,version} code into an installable library that injects its functionality via setuptools entry points | 15:01 |
| bcwaldon | aw man | 15:01 |
| bcwaldon | that would be fantastic | 15:01 |
| mtaylor | bcwaldon: and doesn't trigger the bootstrapping problem of needing to pre-install something before you run setup.py | 15:01 |
| mtaylor | bcwaldon: I'm pretty stoked about it | 15:01 |
| jeblair | chmouel, vishy, bcwaldon: know how to make devstack set 'libvirt_cpu_mode="none"' ? | 15:02 |
| trapni | Hi. I'm having a line `Making asynchronous call on network.cesar1 ... from (pid=14859) multicall` in nova-network.log, and "cesar1" is wrong (changed hostname) - how must I do to teach nova to use the right (new) hostname for nova-network? | 15:02 |
| mtaylor | bcwaldon: in fact, I can even do _more_ now to remove duplication from all of our setup.py files :) | 15:02 |
| *** reidrac has quit IRC | 15:02 | |
| bcwaldon | jeblair: nope | 15:02 |
| * mtaylor does a little dance | 15:02 | |
| bcwaldon | mtaylor: make it happen | 15:02 |
| jeblair | jgriffith: know how to make devstack set 'libvirt_cpu_mode="none"' ? | 15:03 |
| jeblair | any devstack experts in the room willing to help fix it? | 15:03 |
| mtaylor | garyk: we're getting closer... | 15:03 |
| mtaylor | garyk: it _ran_ sphinx that time :) | 15:03 |
| jgriffith | jeblair: nope, not off the top of my head :( | 15:03 |
| russellb | just another line in stack.sh | 15:03 |
| * russellb will look at it | 15:04 | |
| chmouel | jeblair: should not be hard to add | 15:04 |
| jeblair | russellb: thanks! | 15:04 |
| mtaylor | bcwaldon: only blocker I've got right now is that the best name I've been able to come up with for the lib is "versionlib" | 15:04 |
| mtaylor | bcwaldon: which is ridiclous | 15:04 |
| bcwaldon | mtaylor: #agree | 15:04 |
| jeblair | chmouel: it didn't sound hard, but nonetheless, i'm not sure where it goes. :) | 15:04 |
| mtaylor | bcwaldon: but you're good with names -so give me a good lib name and I'll make a release and then start making patches for our stuff to use it | 15:04 |
| chmouel | jeblair: cool I guess russellb is taking a look at it, let me know if you guys need help | 15:05 |
| bcwaldon | mtaylor: grizzly | 15:05 |
| mtaylor | bcwaldon: hahaha | 15:05 |
| jeblair | github suggests: massive-octo-cyril | 15:05 |
| chmouel | prob need to add a CPU_MODE variable or something like that | 15:05 |
| mtaylor | both grizzly and octo-cyril are avail on pypi :) | 15:06 |
| russellb | danpb: should we set it only if libvirt_mode=qemu ? | 15:06 |
| danpb | russellb: no, just set it unconditionally IMHO | 15:06 |
| russellb | err, s/mode/type/ | 15:06 |
| russellb | ok | 15:06 |
| mtaylor | bcwaldon: how ridiculous would it be if openstack grizzly had grizzly as a depend? :) | 15:06 |
| danpb | devstack isn't a case where we really need optimal performance cpus | 15:06 |
| mtaylor | ooh. or I could call it grizzle, just to be different | 15:07 |
| jeblair | mtaylor: perhaps a name more like versionlib? to, like, indicate what it does? seems nicer for a general purpose library... i mean, they didn't call setuptools freezing-robot. | 15:08 |
| bcwaldon | jeblair: no | 15:08 |
| russellb | danpb: jeblair: https://review.openstack.org/9718 | 15:08 |
| danpb | jeblair: to see how close we are to the planes - this is a good example - actually taken standing the ground https://secure.flickr.com/photos/dberrange/4828773030/in/set-72157624422227515/lightbox/ | 15:08 |
| *** sstent has quit IRC | 15:09 | |
| *** mnewby has quit IRC | 15:10 | |
| *** rpedde_away is now known as rpedde | 15:10 | |
| jeblair | russellb: great, thanks! I'll make sure that runs on hpcloud (if it doesn't happen to pick an hpcloud node for its check job) | 15:10 |
| *** sstent has joined #openstack-dev | 15:10 | |
| russellb | jeblair: great thanks | 15:10 |
| *** simo has left #openstack-dev | 15:11 | |
| mtaylor | jeblair: I dunno. kombu and carrot are both fairly popular ... and I'm not sure versionlib actually commuicates that the library not only sets your version from VCS tags, but also injects requirements and sphinx hooks and authors and changelog generation | 15:13 |
| mtaylor | jeblair: I should call it 'libmakemyprojectlikeopenstack' :) | 15:13 |
| *** Gordonz has quit IRC | 15:14 | |
| jeblair | fine. call it grizzly. there is no question who's answer is not grizzly. | 15:14 |
| *** Gordonz has joined #openstack-dev | 15:14 | |
| * jgriffith is changing his nick to grizzly | 15:15 | |
| * jgriffith is changing cinder to grizzly | 15:15 | |
| *** mattray has joined #openstack-dev | 15:16 | |
| *** nati has joined #openstack-dev | 15:16 | |
| *** nati has quit IRC | 15:18 | |
| *** halfss has quit IRC | 15:28 | |
| *** mindpixel has quit IRC | 15:29 | |
| *** armaan has joined #openstack-dev | 15:30 | |
| *** armaan has left #openstack-dev | 15:30 | |
| *** sstent_ has joined #openstack-dev | 15:32 | |
| mtaylor | garyk: ok. one more try | 15:32 |
| *** rnirmal has joined #openstack-dev | 15:32 | |
| jgriffith | jeblair: That cpu_mode setting didn't seem to fix it? (add_nova_opt "libvirt_cpu_mode=none") | 15:33 |
| jgriffith | jeblair: Or is there an associated change that goes with? | 15:33 |
| jeblair | jgriffith: why do you say that? | 15:35 |
| jgriffith | jeblair: Ummm... I patched my stack.sh, ran it, ran exercise.sh and had the same failures | 15:35 |
| jgriffith | jeblair: Did this fix it on the HP Cloud servers? | 15:37 |
| jeblair | ah. i haven't caused it to run on hpcloud yet; the check job ran on rackspace, lemme get it to run on hp. | 15:37 |
| jgriffith | jeblair: Ok, don't know whether I want to hope for it to work or not :) | 15:37 |
| jeblair | jgriffith: https://jenkins.openstack.org/job/gate-tempest-devstack-vm/3452/console | 15:39 |
| *** sstent_ has quit IRC | 15:39 | |
| * jgriffith eyes glued to the console output | 15:41 | |
| *** danpb has quit IRC | 15:41 | |
| jeblair | jgriffith: and it looks like we got a nice, slow instance too. :/ | 15:42 |
| jgriffith | jeblair: Doh... there it goes | 15:42 |
| jgriffith | jeblair: perhaps a good time to go get another cup of coffee | 15:43 |
| *** cp16net|away is now known as cp16net | 15:43 | |
| *** armaan has joined #openstack-dev | 15:43 | |
| *** corXi has quit IRC | 15:43 | |
| *** everett has joined #openstack-dev | 15:45 | |
| *** mattray has quit IRC | 15:47 | |
| *** everett-toews has joined #openstack-dev | 15:48 | |
| *** rohitk has joined #openstack-dev | 15:52 | |
| jeblair | jgriffith: it seems to have addressed the hpcloud problem... | 15:52 |
| jgriffith | jeblair: huh... | 15:53 |
| jgriffith | jeblair: I see, but but but | 15:53 |
| jgriffith | jeblair: Ok, I'm heading down to the apple store now | 15:53 |
| jgriffith | :) | 15:53 |
| *** rods has quit IRC | 15:53 | |
| jeblair | jgriffith: don't do anything rash! :) | 15:53 |
| jeblair | jgriffith: i promise hpcloud isn't running on a stack of macbooks! | 15:54 |
| jgriffith | jeblair: Ahh... well yes, that's a very good point | 15:54 |
| garyk | mtaylor: ++ | 15:54 |
| * jeblair really hopes he's right about that | 15:54 | |
| jgriffith | jeblair: There goes the excuse to my wife for buying the fancy new retina dislay macbook :) | 15:54 |
| garyk | mtaylor: thanks! | 15:54 |
| jgriffith | jeblair: Ok, back to reality. Let me see if I missed somehting on my devstack setup | 15:55 |
| jgriffith | jeblair: I'll start over, and I guess at this point get out of your hair :) | 15:55 |
| jeblair | jgriffith: i hope that works (cause i'm sure out of ideas), and thanks for your help. | 15:56 |
| jgriffith | jeblair: I'm just glad the code reviews should go through smoothly now :) | 15:56 |
| *** mattray has joined #openstack-dev | 15:56 | |
| *** salgado is now known as salgado-lunch | 15:59 | |
| *** arosen has joined #openstack-dev | 16:00 | |
| *** armaan has quit IRC | 16:01 | |
| *** armaan has joined #openstack-dev | 16:01 | |
| *** matwood has joined #openstack-dev | 16:01 | |
| *** mnewby has joined #openstack-dev | 16:05 | |
| *** rods has joined #openstack-dev | 16:07 | |
| *** PotHix has joined #openstack-dev | 16:08 | |
| garyk | mtaylor: thanks! did jenkins merge? | 16:11 |
| *** mnewby has quit IRC | 16:12 | |
| mtaylor | garyk: yah | 16:12 |
| arosen | markmcclain: Hey, I have a quick question about your comment regarding add_nova_opt "quantum_admin_tenant_name=$SERVICE_TENANT_NAME" | 16:13 |
| arosen | I'm trying to reproduce this now. The errors you get are in n-api ? | 16:14 |
| markmcclain | I'll get errors in n-cpu | 16:14 |
| garyk | mtaylor: thanks | 16:15 |
| *** e1mer has quit IRC | 16:15 | |
| *** jerdfelt has quit IRC | 16:19 | |
| *** jerdfelt has joined #openstack-dev | 16:19 | |
| *** matwood has quit IRC | 16:20 | |
| heckj | dolphm, ayoung: quick review: https://review.openstack.org/#/c/9723/ | 16:20 |
| *** matwood has joined #openstack-dev | 16:20 | |
| *** garyk has quit IRC | 16:20 | |
| ayoung | heckj, looking | 16:20 |
| ayoung | heckj, done | 16:22 |
| heckj | thx! | 16:22 |
| *** andrewsmedina has quit IRC | 16:27 | |
| arosen | markmcclain: is this what you are getting? http://codepad.org/EzklhIeV | 16:28 |
| markmcclain | arosen: yes | 16:30 |
| *** mokas has quit IRC | 16:30 | |
| *** davidha has quit IRC | 16:31 | |
| *** jtran has joined #openstack-dev | 16:32 | |
| ttx | TZ reminder: today is bug squashing day, join us on #openstack-bugday if you participate to the effort: http://wiki.openstack.org/BugDays/20120712BugSquashing | 16:33 |
| *** rohitk has quit IRC | 16:40 | |
| *** epim_ has joined #openstack-dev | 16:44 | |
| *** epim has quit IRC | 16:45 | |
| *** epim_ is now known as epim | 16:45 | |
| *** johnpur has joined #openstack-dev | 16:47 | |
| *** ChanServ sets mode: +v johnpur | 16:47 | |
| *** garyk has joined #openstack-dev | 16:48 | |
| *** danpb has joined #openstack-dev | 16:48 | |
| *** m4xmr has quit IRC | 16:48 | |
| *** dhellmann has quit IRC | 16:49 | |
| *** markmcclain has quit IRC | 16:49 | |
| *** dachary has joined #openstack-dev | 16:50 | |
| *** armaan has quit IRC | 16:51 | |
| *** torgomatic has joined #openstack-dev | 16:55 | |
| *** jdurgin has joined #openstack-dev | 16:59 | |
| *** Mandell has joined #openstack-dev | 17:00 | |
| bcwaldon | jeblair: can we take the hp nodes out of rotation? | 17:01 |
| jeblair | bcwaldon: can we merge https://review.openstack.org/#/c/9718/ ? | 17:01 |
| bcwaldon | jeblair: yes | 17:01 |
| bcwaldon | jeblair: approved | 17:02 |
| jeblair | bcwaldon: thanks! | 17:02 |
| *** rohitk has joined #openstack-dev | 17:03 | |
| *** danwent has quit IRC | 17:05 | |
| *** cp16net is now known as cp16net|away | 17:08 | |
| *** cp16net|away is now known as cp16net | 17:08 | |
| *** cp16net is now known as cp16net|away | 17:10 | |
| *** cp16net|away is now known as cp16net | 17:10 | |
| *** m4xmr has joined #openstack-dev | 17:12 | |
| *** danpb has quit IRC | 17:12 | |
| *** m4xmr has quit IRC | 17:12 | |
| *** derekh has quit IRC | 17:14 | |
| *** renier has quit IRC | 17:21 | |
| jtran | when raising an exception in nova-manage, should I just raise Exception('msg'), or do I do a print _('msg') ; raise ? | 17:22 |
| bcwaldon | jtran: is there a precedent in that file? | 17:23 |
| *** jaypipes is now known as jaypipes-afk | 17:23 | |
| jtran | lots of :) | 17:23 |
| bcwaldon | jtran: of both? | 17:23 |
| jtran | yeah and others not listed | 17:23 |
| *** rohitk has left #openstack-dev | 17:24 | |
| jtran | the thing is when just do Exception('msg') , i get nothing returned other than 'check log for errors' but afaik nova-manage doesn't log any errors (anymore) | 17:24 |
| bcwaldon | jtran: well, whatever gets the appropriate message out is fine with me | 17:25 |
| jtran | cool cool | 17:25 |
| dolphm | heckj: typo in https://review.openstack.org/#/c/9723/ | 17:29 |
| *** salgado-lunch is now known as salgado | 17:29 | |
| heckj | dolphm: thanks | 17:29 |
| heckj | dolphm: fixed and updated | 17:31 |
| jgriffith | davidkranz: Ok, we'll see if that helps or makes things worse | 17:33 |
| dolphm | heckj: +2'd | 17:35 |
| *** darraghb has quit IRC | 17:36 | |
| *** gyee has joined #openstack-dev | 17:38 | |
| davidkranz | jgriffith: What do you mean by "that"? | 17:39 |
| *** tongli has joined #openstack-dev | 17:40 | |
| jgriffith | davidkranz: You suggested an email... | 17:40 |
| davidkranz | jgriffith: Yes, but I didn't see one. | 17:40 |
| jgriffith | davidkranz: I just sent on eout | 17:40 |
| jgriffith | davidkranz: Hmmm... | 17:40 |
| andrewbogott | Can someone who knows more than me about python internals explain why this would ever fail? self.assertEqual(cfg.CONF.__getattr__('notification_driver'), cfg.CONF.notification_driver) | 17:42 |
| jgriffith | davidkranz: Let me know if you don't see it in the next few minutes, was sent 8 minutes ago | 17:43 |
| davidkranz | jgriffith: k | 17:43 |
| *** mokas has joined #openstack-dev | 17:45 | |
| andrewbogott | ...because it is failing for me. __getattr__ is overloaded in clever ways, but apparently just accessing .members bypasses __getattr__ which... I thought the whole point of __getattr__ was that it was called during .access. | 17:46 |
| bcwaldon | andrewbogott: what's the class of cfg.CONF? | 17:46 |
| *** thingee_zzz is now known as thingee | 17:46 | |
| bcwaldon | andrewbogott: nvm, so looking through the code I don't see anything obvious | 17:47 |
| andrewbogott | openstack.common.cfg.CommonConfigOpts | 17:47 |
| bcwaldon | andrewbogott: maybe you should determine if they're following the same path through ConfigOpts | 17:47 |
| andrewbogott | I have a test that's failing due to an attempt to override a cfg variable... but when I access that cfg var later (using a simple .reference) the override is not present. | 17:48 |
| bcwaldon | the two method of accessing | 17:48 |
| bcwaldon | andrewbogott: two different tests? | 17:48 |
| andrewbogott | bcwaldon: I litter cfg with debug lines... I don't see evidence that anything in cfg is called with the .access. | 17:49 |
| andrewbogott | Which maybe is expected? But that means that accessing config values that way will never work right. | 17:49 |
| bcwaldon | andrewbogott: is this a test | 17:49 |
| andrewbogott | bcwaldon: yes... I have it narrowed down so that I can get an immediate failure in a setUp() function: | 17:50 |
| andrewbogott | self.config( | 17:50 |
| andrewbogott | notification_driver='openstack.common.notifier.no_op_notifier') | 17:50 |
| andrewbogott | self.assertEqual(cfg.CONF.__getattr__('notification_driver'), cfg.CONF.notification_driver) | 17:50 |
| bcwaldon | andrewbogott: ...as in, is something getting stubbed out in a test | 17:50 |
| andrewbogott | bcwaldon: Yes, stubbing behavior is exactly what I'm trying to fix and/or understand. | 17:52 |
| bcwaldon | andrewbogott: so are you sure that cfg.CONF isn't some fake object? | 17:52 |
| bcwaldon | andrewbogott: you can also ensure they're the *same* object | 17:52 |
| andrewbogott | I'm pretty much the only one who's using self.config() in common, and I'm finding that its behavior is inconsistent. Whereas if I use self.stubs.Set(cfg.CONF, 'notification_driver', etc.) it does what I expect. | 17:53 |
| andrewbogott | bcwaldon: In the above assertEqual example... how could they be different objects? (And, in the __getattr__ case, my debug lines are traversed.) | 17:54 |
| bcwaldon | andrewbogott: dono, I guess I won't be much help here | 17:55 |
| *** cp16net is now known as cp16net|away | 17:55 | |
| andrewbogott | bcwaldon: Do you share my expectation __getattr__ is called when you access a member with .? | 17:55 |
| bcwaldon | andrewbogott: yes, that's a fact | 17:55 |
| andrewbogott | huh | 17:56 |
| bcwaldon | andrewbogott: well, it might be __get__, not __getattr__ | 17:56 |
| bcwaldon | andrewbogott: I need to brush up on the difference | 17:56 |
| bcwaldon | andrewbogott: ah, __getattr__ is only called if no class attribute exists | 17:56 |
| bcwaldon | andrewbogott: so calling it directly might return a different response that whatever is actually set as an attribute | 17:57 |
| bcwaldon | andrewbogott: try the assert with __get__ instead | 17:57 |
| andrewbogott | If I call __get__ I get 'no such option' | 17:57 |
| andrewbogott | But, that might be a clue of some sort. | 17:57 |
| bcwaldon | yes | 17:58 |
| bcwaldon | a clue! | 17:58 |
| andrewbogott | Interesting, it's calling _get with the argument __get__ which is failing. | 17:59 |
| * andrewbogott 's head spins | 17:59 | |
| *** GheRivero has quit IRC | 18:00 | |
| *** armaan has joined #openstack-dev | 18:00 | |
| davidkranz | jgriffith: Haven't seen it... | 18:01 |
| jgriffith | davidkranz: that's weird... | 18:02 |
| jgriffith | anybody get my email? | 18:03 |
| *** mnewby has joined #openstack-dev | 18:04 | |
| *** danwent has joined #openstack-dev | 18:05 | |
| jgriffith | davidkranz: https://lists.launchpad.net/openstack/msg14533.html | 18:06 |
| jgriffith | davidkranz: It made it to the archive at any rate | 18:06 |
| davidkranz | jgriffith: Not sure why I didn't get it. | 18:06 |
| *** cp16net|away is now known as cp16net | 18:06 | |
| *** mdomsch has quit IRC | 18:07 | |
| *** kbringard has joined #openstack-dev | 18:07 | |
| *** mdomsch has joined #openstack-dev | 18:07 | |
| *** dachary has quit IRC | 18:08 | |
| *** salv-orlando has joined #openstack-dev | 18:08 | |
| *** dachary has joined #openstack-dev | 18:08 | |
| *** markmcclain has joined #openstack-dev | 18:11 | |
| vishy | bcwaldon, comstud, pvo, jorge: any idea if there is documentation on instance faults anywhere? | 18:12 |
| vishy | it doesn't look like it is in any of our docs. | 18:12 |
| annegentle | vishy: what are instance faults exactly? | 18:12 |
| vishy | annegentle: when an instance is in error state, there is a potential list of errors that is displayed through the api | 18:12 |
| *** salv-orlando has left #openstack-dev | 18:14 | |
| *** salv-orlando has joined #openstack-dev | 18:14 | |
| *** renier has joined #openstack-dev | 18:14 | |
| *** dhellmann has joined #openstack-dev | 18:15 | |
| *** mnewby has quit IRC | 18:15 | |
| ewindisch | russellb: disabling kombu tests was not intended to be merged. Forgot the -D on git-view | 18:15 |
| zykes- | vishy: you got any clue on what can be write if a IP doesn't get allocated in nova with quantum ? | 18:16 |
| bcwaldon | vishy: yep | 18:17 |
| *** anniec has quit IRC | 18:17 | |
| bcwaldon | vishy: http://docs.openstack.org/api/openstack-compute/2/content/Asynchronous_Faults-d1e2009.html | 18:17 |
| russellb | ewindisch: k, i see that it's a draft now, that's fine | 18:18 |
| russellb | ewindisch: i also left a comment on your other patch about pyzmq in test-requires ... | 18:18 |
| russellb | that won't work for build slaves. | 18:18 |
| russellb | pyzmq will blow up since zeromq won't be installed | 18:18 |
| * russellb went down this road while trying to run your tests in nova ... | 18:19 | |
| russellb | pyzmq-staic (or something like that) looked promising, but I don't think i got it working | 18:19 |
| russellb | that one was supposed to include the base C lib as well | 18:19 |
| ewindisch | russellb: the problem is that the slaves already have pyzmq installed (and working) | 18:20 |
| russellb | o r l y | 18:20 |
| ewindisch | so I wanted it to reflect reality... | 18:20 |
| ewindisch | and pyzmq-static never worked, and is no longer in pypi. | 18:20 |
| russellb | why do they have it installed? just got lucky? | 18:20 |
| ewindisch | apparently. | 18:20 |
| russellb | heh, alrighty then. | 18:20 |
| russellb | but then it will break it for people running tests locally | 18:21 |
| ewindisch | actually, I wonder if that is the problem on the 2.6 tests. Maybe the 2.6 slaves have the pyzmq-static version that is broken | 18:21 |
| clarkb | python-zmq is installed by puppet for something on the slaves and that installs libzmq | 18:21 |
| russellb | clarkb: ah cool, thanks | 18:21 |
| russellb | ewindisch: but for anyone else running the unit tests ... adding that will break things i think | 18:22 |
| *** mikal has quit IRC | 18:23 | |
| russellb | if only we could add it to test-requires as an optional thing and the venv build would just continue and ignore it if it fials | 18:23 |
| ewindisch | russellb: right now, the tests are disabled entirely because jenkins has pyzmq and the tests were failing. | 18:23 |
| * russellb nods | 18:23 | |
| russellb | i'm just talking about your proposed patch | 18:23 |
| ewindisch | right. | 18:23 |
| russellb | annnnnyway | 18:23 |
| *** rnirmal has quit IRC | 18:24 | |
| * russellb keeps digging through jenkins | 18:24 | |
| russellb | errr, gerrit. | 18:24 |
| *** GheRivero has joined #openstack-dev | 18:24 | |
| ewindisch | russellb: I agree it would break local tests. I don't have a recommendation except to maybe look into tox.ini, maybe we can build it ? ;-) | 18:25 |
| russellb | build zeromq? maybe.. | 18:26 |
| jeblair | mtaylor: i think you added python-zmq for the nova zeromq unittests, right? | 18:26 |
| *** mikal has joined #openstack-dev | 18:27 | |
| russellb | if you're doing that, installing python-qpid would rock :-) ... it's not in pypi :( | 18:27 |
| russellb | happy and sad in the same line, i'm so emo | 18:27 |
| jeblair | ewindisch: can you elaborate on "pyzmq-static version that is broken" | 18:31 |
| vishy | russellb: https://review.openstack.org/#/c/8329/ | 18:32 |
| jeblair | ewindisch: the only zmq related packages on the oneiric sslaves are | 18:32 |
| jeblair | libzmq1 2.1.9-1 | 18:32 |
| andrewbogott | bcwaldon: Here is a summation of my confusion, in handy patch form: https://review.openstack.org/#/c/9731/ | 18:32 |
| jeblair | python-zmq 2.1.9-1 | 18:32 |
| vishy | russellb: thoughts on the rpcapi modifications? | 18:32 |
| andrewbogott | bcwaldon: (Probably I just have to wait for markmc to get back and explain.) | 18:32 |
| bcwaldon | andrewbogott: ha, ok | 18:33 |
| ewindisch | jeblair: referring to a broken pypi package that no longer exists. Just making sure that wasn't installed. | 18:33 |
| russellb | vishy: will look, thanks for pointing it out | 18:33 |
| ewindisch | jeblair: however, I'm quite sure that 2.1.9 doesn't work with the zeromq driver. I'm almost certain that is the problem. I know that 2.1.11 works, which is what is running on the precise boxes. | 18:33 |
| *** pixelbeat has quit IRC | 18:34 | |
| *** AlanClark has joined #openstack-dev | 18:34 | |
| ewindisch | jeblair: what I'll do is put in a test for the zeromq version in the tests. They'll skip if pyzmq is missing, or is too low a version. | 18:34 |
| clarkb | ewindisch: can you pin the pyzmq package to version 2.1.9 to test? | 18:34 |
| *** cmagina has quit IRC | 18:34 | |
| ewindisch | clarkb: I could. I've tested it in the past, and it failed. | 18:34 |
| *** rnirmal has joined #openstack-dev | 18:34 | |
| jeblair | ewindisch: for reference, jenkins slave configuration is here: https://github.com/openstack/openstack-ci-puppet/blob/master/modules/jenkins_slave/manifests/init.pp | 18:35 |
| ewindisch | cool. | 18:36 |
| *** cmagina has joined #openstack-dev | 18:36 | |
| mtaylor | jeblair: yeah | 18:36 |
| ewindisch | This is what I propose to address these concerns: | 18:36 |
| ewindisch | * skip tests on pyzmq < 2.1.11 (tested-working); I can do 2.1.9 testing later and see if this can be fixed. | 18:37 |
| dprince | bcwaldon: Is there a particular reason you didn't +2 this https://review.openstack.org/#/c/9661/? Your concern was about the 500 right? | 18:38 |
| ewindisch | * add cdeps building to tox to make the tests self-contained on c-based depends. This will also help us with qpid | 18:38 |
| bcwaldon | dprince: I wanted to leave the discussion open, but then I went and saw the bug/fix in keystone you filed | 18:38 |
| bcwaldon | dprince: I'll +2, just didnt think to come back to it | 18:39 |
| dprince | bcwaldon: Reason I ask is I'd like to do the same thing to my Glance branch... which is blocking my entire patch series. | 18:39 |
| *** alaski has quit IRC | 18:39 | |
| bcwaldon | dprince: do you want to i18n that error message? | 18:39 |
| bcwaldon | dprince: well, would you please :) | 18:39 |
| dprince | bcwaldon: Okay. I'll make the same change in Glance too then. | 18:39 |
| bcwaldon | dprince: yep, will +2 after that | 18:39 |
| jeblair | mtaylor: you're the tox expert, what do you think about ewindisch's suggestions ^ ? | 18:41 |
| *** andrewsmedina has joined #openstack-dev | 18:41 | |
| *** mokas has quit IRC | 18:41 | |
| *** mokas has joined #openstack-dev | 18:41 | |
| *** mikal has quit IRC | 18:42 | |
| *** mikal has joined #openstack-dev | 18:43 | |
| mtaylor | looking | 18:43 |
| mtaylor | hrm. | 18:43 |
| mtaylor | what's cdeps? | 18:43 |
| ewindisch | https://review.openstack.org/9733 | 18:44 |
| mtaylor | ewindisch: that's a draft, you'll need to add me as a reviewer | 18:44 |
| ewindisch | added | 18:45 |
| ewindisch | This isn't tested and isn't fully baked - just a POC for review before I go down this rabbit hole | 18:46 |
| mtaylor | yeah - so, there's a few things where this will go to the bad place | 18:47 |
| mtaylor | (not saying no, just things that will need to be solved for it to be usable) | 18:47 |
| mtaylor | cloning from github is pretty much always a terrible idea, as it fails more than you'd expect when it's part of our automated testing | 18:47 |
| bcwaldon | mtaylor: learning is fun | 18:48 |
| mtaylor | we do everything in our power to avoid it (and any other network resource) as much as possible a | 18:48 |
| ewindisch | mtaylor: clearly, we shouldn't install the c-dependencies globally. The C include paths will need to be passed to pip | 18:48 |
| mtaylor | ewindisch: no, I mean your cdeps build does a clone | 18:48 |
| mtaylor | that will wail | 18:48 |
| mtaylor | fail | 18:48 |
| ewindisch | mtaylor: yes, I know. I was typing that out when you mentioned the clone | 18:48 |
| mtaylor | :) | 18:49 |
| ewindisch | yeah, the clone is an issue... | 18:49 |
| mtaylor | so, we can grab release tarballs | 18:49 |
| mtaylor | because we could potentially mirror those | 18:49 |
| mtaylor | but then we've got additional problems, in that we'd need to verify for each of these things that they're installable into the virtualenv without installing portions of them globally | 18:50 |
| ewindisch | which is still requiring network access | 18:50 |
| mtaylor | network access to _our_ servers is fine | 18:51 |
| ewindisch | sure, C code has dependencies too, which we'd have to apt-get (or yum, or worse) | 18:51 |
| ewindisch | and /that/ is a rabbit-hole | 18:51 |
| mtaylor | we don't tend to lose network access to our own servers very often | 18:51 |
| mtaylor | yup | 18:51 |
| mtaylor | we tried a version of part of this a while ago when we still based builds off of apt packages | 18:51 |
| mtaylor | in that we had a ppa and we keep custom or specific versions of debs in that ppa | 18:51 |
| mtaylor | but then we were starting to create a little mini-distro | 18:52 |
| mtaylor | I like the first part of your suggestion - which is to put in the version check in the tests... | 18:52 |
| mtaylor | then I think the next step from a moving-forward perspective is to bug zul to get an updated version of the libs into ubuntu | 18:52 |
| zul | what libs? | 18:53 |
| ewindisch | mtaylor: so the issue here isn't for the automated tests, in my opinion. | 18:53 |
| ewindisch | mtaylor: we can assure those libraries exist via the slave manifest | 18:53 |
| ewindisch | we're having an issue that even if that stuff exists in the slave's manifest, how do we make sure the tests pass on individual user's machines? | 18:54 |
| mtaylor | right. which is why we've been really reluctant to install non-standard anything on the slaves | 18:54 |
| *** davidha has joined #openstack-dev | 18:54 | |
| mtaylor | (I hate all of the solutions to this problem, btw) | 18:55 |
| ewindisch | If I add pyzmq to tests-requires, the tests will fail on individual user's machines. | 18:55 |
| ewindisch | if I don't add it to tests-requires, but install pyzmq via the manifest, the jenkins slaves are going to run tests that weren't run on the developer's machine | 18:55 |
| mtaylor | wait - is pyzmq installable via pip? | 18:56 |
| mtaylor | I thought it was one of those swig-based libs that wasn't | 18:56 |
| ewindisch | mtaylor: yes, but not the c-dependency that the pip-package requires. | 18:56 |
| ewindisch | it is Cython based, by the way. | 18:56 |
| clarkb | libzmq1 on ubuntu I think | 18:56 |
| mtaylor | ah. so it's the libzmq1 dep that's the issue? | 18:57 |
| mtaylor | is the precise version of libzqm1 sufficient? | 18:57 |
| *** lorin1 has quit IRC | 18:57 | |
| mtaylor | (sorry for being seemingly daft - bear with me for a sec) | 18:57 |
| *** armaan has left #openstack-dev | 18:57 | |
| jeblair | mtaylor: isn't the question: "is the oneiric version of libzmq1 sufficient for pip pyzmq?" | 18:58 |
| mtaylor | jeblair: AH | 18:58 |
| ewindisch | mtaylor: no, the precise version of libzmq1 is not sufficient. | 18:58 |
| mtaylor | ok. well, that makes it slightly easier | 18:59 |
| ewindisch | or oneiric. damn i keep foregetting the names | 18:59 |
| ewindisch | and I just added an 'e' where it didn't belong. | 18:59 |
| *** dprince has quit IRC | 18:59 | |
| mtaylor | what version of libzmq1 is sufficient? | 18:59 |
| ewindisch | 2.1.11 ; This is in precise, but not in oneiric. | 18:59 |
| jeblair | ah, you said that earlier, i thought that was a pyzmq version # | 19:00 |
| clarkb | jeblair: they follow each other to make interversion dependencies easy to sort out | 19:01 |
| clarkb | s/interversion/interproject/ | 19:01 |
| *** thingee has quit IRC | 19:01 | |
| mtaylor | ok. so the situation we find ourselves in is this: we cannot support pyzmq as a dependency on oneiric | 19:01 |
| *** japage2 has joined #openstack-dev | 19:02 | |
| mtaylor | so how the _hell_ do we express that in a meaningful manner | 19:02 |
| *** jog0 has joined #openstack-dev | 19:02 | |
| mtaylor | ewindisch: what happens if you pip install pyzmq 2.1.11 on oneiric with libzmq 2.1.9 ... does it blow up? or does it work in a degraded manner? | 19:02 |
| ewindisch | mtaylor: I can have the tests simply skip with <2.1.11, that solves that problem, imho. | 19:03 |
| mtaylor | it _might_ solve it | 19:03 |
| *** japage has quit IRC | 19:03 | |
| *** alrs has joined #openstack-dev | 19:03 | |
| mtaylor | the problem is, we'd put the pip-requires at pyzmq>2.1.11 ... but pip on oneiric _is_ going to be able to find that | 19:03 |
| ewindisch | mtaylor: it degrades… but it will work as if you had pyzmq 2.1.9 installed (minus any bug fixes in the python) | 19:04 |
| mtaylor | ewindisch: AHA! | 19:04 |
| mtaylor | cool. that's stellar | 19:04 |
| mtaylor | can you make your 2.1.11 check fail on pyzmq 2.1.11 /libzmq1 2.1.9 ? | 19:04 |
| *** thingee has joined #openstack-dev | 19:04 | |
| ewindisch | maybe | 19:04 |
| mtaylor | if you can do that then we're golden | 19:05 |
| ewindisch | seems so zmq.core.version.zmq_version() | 19:05 |
| mtaylor | if that's not possible, then I think what we want to do is put the zmq tests into the test-options file and run them as part of the "full" env | 19:05 |
| mtaylor | PERFECT | 19:05 |
| mtaylor | that tests for the capability we need, will mean that the test suite will properly skip on oneiric and run on precise and a dev can reasonably be expected to deal with it | 19:06 |
| *** gyee has quit IRC | 19:06 | |
| ewindisch | mtaylor: the fact remains that we can't put pyzmq into test-requires though | 19:06 |
| mtaylor | why? | 19:07 |
| ewindisch | because tox will bail if libzmq1 isn't intalled. | 19:07 |
| mtaylor | ah- that's fine - put it into test-options | 19:07 |
| mtaylor | that's the one that only gets pulled in on a "full" test run - and where we put things that expect the dev to have extra crap installed | 19:07 |
| ewindisch | do we need to create that? | 19:07 |
| mtaylor | it's there, as is the full env | 19:07 |
| ewindisch | (where is that file?) | 19:07 |
| mtaylor | tools | 19:07 |
| mtaylor | we do need to make the jenkins job to run that env | 19:08 |
| ewindisch | mtaylor: This is in openstack-common. | 19:08 |
| mtaylor | which I can do today | 19:08 |
| ewindisch | maybe you're talking about Nova or such? | 19:08 |
| mtaylor | ewindisch: ah- then we just need to copy that facility in from nova | 19:08 |
| mtaylor | ewindisch: make a tools/test-options file | 19:08 |
| ewindisch | ah, yeah, I see it now. (looks at nova) | 19:08 |
| mtaylor | and grab the [testenv:full] section as well | 19:08 |
| ewindisch | yep, I see it. | 19:09 |
| mtaylor | and I'll work on getting jenkins jobs added for nova and openstack-common | 19:09 |
| mtaylor | (we have a similar issue with MySQL-python and libvirt) | 19:09 |
| *** datsun180b has quit IRC | 19:10 | |
| ewindisch | +1 thanks monty | 19:10 |
| ewindisch | russellb: sorry, I'm not sure this helps you in the short-term. | 19:11 |
| *** mnewby has joined #openstack-dev | 19:14 | |
| *** jakedahn_zz is now known as jakedahn | 19:15 | |
| *** littleidea has joined #openstack-dev | 19:15 | |
| mtaylor | we'll get something going :) | 19:17 |
| zul | jgriffith: ping | 19:19 |
| jgriffith | zul: poing | 19:19 |
| jgriffith | poing | 19:19 |
| jgriffith | pong!!!!! | 19:19 |
| maoy | vishy: there seems to be a confusing use between InstanceNotFound and NotFound in the libvirt driver | 19:20 |
| zul | jgriffith: so nova-volumes/cinder right now doesnt save the iscsi volumes if you reboot the server so we were discussing this internally and we were thinking of switching to use the tgt configuration files, so when the server reboots the volumes are still there | 19:20 |
| maoy | vishy: sth like throw one and catch the other | 19:21 |
| maoy | vishy: what's the design behind the two exceptions? which one is preferred? thx. | 19:21 |
| jgriffith | zul: You mean just in your setup, or you mean proposing a blueprint? | 19:22 |
| zul | jgriffith: proposing a blueprint | 19:22 |
| maoy | vishy: scratch that. never mind. :) | 19:23 |
| zul | i was just wanting to get your opinon | 19:23 |
| jgriffith | zul: There were some reasons for not using tgt as I recall | 19:23 |
| jgriffith | zul: But I think putting it out there for everybody to review is the best thing to do | 19:23 |
| zul | jgriffith: right but iscsitarget would have the same problem | 19:23 |
| jgriffith | zul: Understood | 19:24 |
| zul | jgriffith: since it isnt persistence across reboots | 19:24 |
| jgriffith | I'd submit the blueprint | 19:24 |
| *** littleidea has quit IRC | 19:25 | |
| zul | jgriffith: ok i have been toying around with something as well locally | 19:26 |
| jgriffith | zul: Yeah? | 19:26 |
| ewindisch | mtaylor: actually, all of that said, I actually do have the tests passing on MacOS with 2.1.7… /sigh/ | 19:27 |
| zul | jgriffith: yep | 19:27 |
| jgriffith | zul: lay it on me sometime, always like to hear others ideas | 19:27 |
| *** novas0x2a|laptop has joined #openstack-dev | 19:27 | |
| ewindisch | I will be building a Oneiric box to test this all out. At least for now, I'll have something to commit and pass tests. | 19:27 |
| ewindisch | (on the slaves) | 19:28 |
| zul | jgriffith: basically tgt can use other directories other than /etc/tgt so basically when the volume is created you would create the tgt configuration file in something like /var/lib/nova/volumes and remove it when the volume goes away | 19:28 |
| jgriffith | zul: We should talk sometime on that, I had a similar idea but put it on the back burner until we're ready to work on things like multiple back-ends | 19:30 |
| zul | jgriffith: ok ill propose a blueprint at least | 19:30 |
| jgriffith | zul: cool! Thanks | 19:30 |
| vishy | maoy: figured it out? | 19:32 |
| *** mokas has quit IRC | 19:32 | |
| ogelbukh | jgriffith: zul: is it possible to make recreating exports part of nova-volume/cinder service initialization? | 19:32 |
| zul | ogelbukh: possibly | 19:33 |
| jgriffith | ogelbukh: zul: that's not something that would trivial I don't think | 19:33 |
| jgriffith | ogelbukh: But ANYTHING is possible :) | 19:34 |
| zul | jgriffith/ogelbukh: what about like a scheduled task that just polls the running volumes and updates the configuration file | 19:35 |
| jog0 | vishy: I am looking for reviews for my rebased aggregates part 1 patch | 19:35 |
| jgriffith | zul: don't know how well that would scale | 19:35 |
| jgriffith | zul: polling is kind of a dirty word | 19:35 |
| zul | jgriffith: right | 19:36 |
| *** liemmn has joined #openstack-dev | 19:36 | |
| *** liemmn has left #openstack-dev | 19:37 | |
| jgriffith | zul: ogelbukh: I like the idea, would like to see some thought put into it and figure out what we can do | 19:39 |
| *** zaitcev has joined #openstack-dev | 19:40 | |
| rnirmal | the existing code does do an ensure_export at volume service startup | 19:41 |
| rnirmal | not sure all drivers have implemented that | 19:41 |
| *** johnpur has quit IRC | 19:42 | |
| *** japage2 has quit IRC | 19:47 | |
| ttx | Last reminder of the day: today is bug squashing day, join us on #openstack-bugday if you participate to the effort: http://wiki.openstack.org/BugDays/20120712BugSquashing | 19:47 |
| jog0 | vishy: thanks | 19:48 |
| zul | rnirmal: i just did a quick check and only one of the drivers do that | 19:51 |
| rnirmal | well there you go, I remembered the manager did something on startup | 19:51 |
| rnirmal | zul: it would be great to get that updated for all the drivers that support it | 19:52 |
| *** jaypipes-afk is now known as jaypipes | 19:52 | |
| rnirmal | support/require it | 19:52 |
| *** armaan has joined #openstack-dev | 19:53 | |
| *** japage has joined #openstack-dev | 19:56 | |
| zul | rnirmal: im surprised we arent storing the luns in the table that would make it easier as well i think | 19:56 |
| rnirmal | the lun information is stored in provider_location in the volumes table, it's hacky at best, but here's the layout for that column <ip> <iqn> <lun> | 19:57 |
| rnirmal | zul: ^ | 19:58 |
| rnirmal | and that's also very dependent on the driver, it's what the driver provides, so not all drivers may have it | 19:59 |
| *** sandywalsh has quit IRC | 20:00 | |
| zul | rnirmal: heh i only care about tgt right now because thats what ubuntu uses | 20:01 |
| jgriffith | vishy: Do you know anything about the status of keystone/quotas (https://blueprints.launchpad.net/keystone/+spec/store-quota-data) | 20:01 |
| jgriffith | vishy: Trying to decide how to proceed with Cinder/Quotas for now | 20:01 |
| heckj | jgriffith: I doubt it | 20:01 |
| vishy | jgriffith: nope | 20:02 |
| heckj | :-) | 20:02 |
| jgriffith | heckj: vishy: It was worth a shot :) | 20:02 |
| vishy | jgriffith: just port the quota code over from nova for now | 20:02 |
| jaypipes | jgriffith, vishy: wow, talk about making a fool of oneself. | 20:02 |
| heckj | jgriffith: Everett started the work in a fork on a local github repo, but hasn't submited anything for review as yet to enable on the keystone side | 20:02 |
| jgriffith | heckj: yeah, I sent him a note the other day when you and I talked | 20:02 |
| jgriffith | heckj: I'll move forward with porting it | 20:03 |
| everett-toews | heckj: jgriffith: i ran into some issues with my first cut. need to retool and try again. | 20:05 |
| heckj | everett-toews: sounds good. | 20:06 |
| jgriffith | everett-toews: ahh... hey there. I'll just keep a look out for things and port as vishy said for now | 20:06 |
| jgriffith | everett-toews: Thanks for letting me know though! | 20:06 |
| heckj | jgriffith: ++ | 20:06 |
| everett-toews | jgriffith: i'll try to have something working by the end of this week and i'll share the link to the branch with you. | 20:07 |
| *** BLZbubba_ has quit IRC | 20:08 | |
| jgriffith | everett-toews: Great, thanks! | 20:08 |
| *** ncode has quit IRC | 20:08 | |
| *** BLZbubba has joined #openstack-dev | 20:09 | |
| *** andrewsmedina has quit IRC | 20:12 | |
| *** mokas has joined #openstack-dev | 20:13 | |
| dolphm | heckj: i put /policies/ stuff out there for service & client on existing /v2.0/ but i'm thinking both should wait to land in /v3/, no? | 20:17 |
| jaypipes | jgriffith: that was an excellent followup email. thx for doing that. | 20:17 |
| heckj | dolphm: 3.0 is what I was thinking - there's no formal API, and we'd be just making an extension to 2.0 for it | 20:18 |
| jgriffith | jaypipes: Thanks! Hope it didn't just get lost in the noise | 20:18 |
| jaypipes | jgriffith: well, nothing you can do about that! | 20:18 |
| jgriffith | jaypipes: Nope | 20:18 |
| dolphm | heckj: i know rax wants rbac in 2.0 as an extension anyway, based on whatever lands in 3.0 | 20:18 |
| *** rafaduran has joined #openstack-dev | 20:19 | |
| heckj | OK by me - I'm preferring to focus on forward, but I get the desire. I think the fastest way will be to implement 3 and get it stable and solid | 20:19 |
| dolphm | heckj: agree | 20:20 |
| *** ncode has joined #openstack-dev | 20:21 | |
| *** Ryan_Lane has joined #openstack-dev | 20:22 | |
| *** armaan has left #openstack-dev | 20:23 | |
| *** jaypipes has quit IRC | 20:26 | |
| *** littleidea has joined #openstack-dev | 20:32 | |
| *** steveb___ has joined #openstack-dev | 20:36 | |
| *** datsun180b has joined #openstack-dev | 20:38 | |
| ewindisch | russellb: I acknowledge the qpid-devstack thing is not a bug. The question is: should we approve and merge the change? I think it would be good to be consistent. | 20:38 |
| russellb | all of the docs everywhere still reference the old syntax | 20:42 |
| russellb | so i don't think it's important to change | 20:42 |
| russellb | and making the option value even longer in all the docs doesn't seem ideal | 20:42 |
| russellb | if we were going to recommend something different, i'd rather add support for shorthand | 20:43 |
| russellb | 'qpid', 'kombu', 'zeromq' | 20:43 |
| *** ncode has quit IRC | 20:43 | |
| *** ncode has joined #openstack-dev | 20:43 | |
| *** ncode has joined #openstack-dev | 20:43 | |
| ewindisch | russellb: that won't work too well if someone wants to use a driver outside of openstack-common (or nova) | 20:45 |
| russellb | well we'd support that too, but that'd be advanced / very uncommon usage | 20:46 |
| russellb | for the general user, a python path doesn't make any sense | 20:46 |
| russellb | it's ugly and not user friendlyt | 20:46 |
| russellb | nova.rpc.impl_whatever is bad enough, nova.openstack.common.rpc.impl_whatever is worse | 20:46 |
| dolphm | ayoung: what's this reverify/recheck voodoo? | 20:49 |
| ewindisch | I guess the only way we could do that is to do string parsing. a short string expands to the openstack.common path, but a string starting with some token gets the token stripped and the module is used directly | 20:49 |
| ayoung | dolphm, it retriggers the Jenkins jobs | 20:49 |
| dolphm | ayoung: which one? | 20:49 |
| ayoung | recheck is for the job run on initial upload of the patch | 20:49 |
| ayoung | I made the mistake of running reveirfy | 20:50 |
| ewindisch | then, "kombu" would load nova.rpc.kombu, while ":kombu" would load 'kombu' from PYTHONPATH | 20:50 |
| russellb | or just check to see if we have a mapping for what is specified, otherwise assume it's a path to import directly | 20:50 |
| ayoung | reverify should only be done after approval | 20:50 |
| dolphm | ayoung: hmm, cool -- thanks! | 20:50 |
| *** jaypipes has joined #openstack-dev | 20:50 | |
| ayoung | But that is what I am here for: making the embarrassing mistakes so you don't have to. | 20:50 |
| clarkb | running reverify before approvals won't cause merges to happen but it will potentially slow down gate tests for changes that have been approved | 20:51 |
| ewindisch | russellb: I personally think that is dirtier. It isn't very deterministic. | 20:51 |
| dolphm | clarkb: is reverify higher priority than recheck or something? | 20:52 |
| russellb | seems fine to me. *shrugs* | 20:52 |
| ayoung | dolphm, I resent the email to you, minus the [CI] tag that probably filtered it into an unread message box | 20:53 |
| dolphm | ayoung: haha thanks | 20:53 |
| dolphm | ayoung: reading | 20:53 |
| clarkb | dolphm: no, but gate tests are run in a dependent queue and if you get into that queue unnecessarily everything after you has been slowed down | 20:54 |
| dolphm | clarkb: gotcha | 20:55 |
| *** GheRivero has quit IRC | 20:56 | |
| *** japage has quit IRC | 20:58 | |
| *** lts has quit IRC | 20:58 | |
| *** rpedde is now known as rpedde_away | 21:05 | |
| *** Ryan_Lane has quit IRC | 21:06 | |
| *** tongli has quit IRC | 21:07 | |
| *** maoy has quit IRC | 21:07 | |
| *** Ryan_Lane has joined #openstack-dev | 21:11 | |
| *** matwood_ has joined #openstack-dev | 21:11 | |
| *** AlanClark has quit IRC | 21:12 | |
| *** mdomsch has quit IRC | 21:14 | |
| *** matwood has quit IRC | 21:15 | |
| *** matwood_ has quit IRC | 21:16 | |
| *** littleidea has quit IRC | 21:17 | |
| *** salgado has quit IRC | 21:17 | |
| *** anniec has joined #openstack-dev | 21:17 | |
| *** anniec_ has joined #openstack-dev | 21:18 | |
| vishy | jgriffith: we still need a change that switches the gating on the other projects to use cinder | 21:19 |
| jgriffith | vishy: Other than the two devstack gate jobs? | 21:20 |
| jgriffith | vishy: or changes rather? | 21:20 |
| jgriffith | vishy: https://review.openstack.org/#/c/9747/1 | 21:21 |
| *** anniec has quit IRC | 21:22 | |
| *** anniec_ is now known as anniec | 21:22 | |
| *** markvoelker has quit IRC | 21:22 | |
| vishy | jgriffith: aye | 21:22 |
| vishy | jgriffith: devstack-vm-gate.sh, changing enabled services there | 21:22 |
| jgriffith | vishy: It's submitted: https://review.openstack.org/#/c/9746/ | 21:23 |
| jgriffith | vishy: Unless I'm missing something else? | 21:23 |
| *** rnirmal has quit IRC | 21:24 | |
| *** Ryan_Lane has quit IRC | 21:25 | |
| *** arosen1 has joined #openstack-dev | 21:25 | |
| vishy | jgriffith: ah thanks, i missed that one | 21:26 |
| jgriffith | vishy: cool | 21:26 |
| *** datsun180b has quit IRC | 21:28 | |
| *** dubsquared has quit IRC | 21:32 | |
| *** anniec has quit IRC | 21:36 | |
| *** armaan has joined #openstack-dev | 21:38 | |
| *** dhellmann has quit IRC | 21:40 | |
| *** salv-orlando has quit IRC | 21:40 | |
| *** arosen1 has quit IRC | 21:41 | |
| *** rkukura has quit IRC | 21:43 | |
| *** arosen1 has joined #openstack-dev | 21:44 | |
| *** Dr_Who has quit IRC | 21:46 | |
| vishy | jog0: if you put a joinedload('hosts') in aggregate_get it will fix lazy load error | 21:47 |
| *** arosen2 has joined #openstack-dev | 21:48 | |
| vishy | jog0: much better than the hack you have in there | 21:48 |
| *** arosen1 has quit IRC | 21:50 | |
| jog0 | vishy: I will change it now | 21:54 |
| bcwaldon | jk0: would you like to review a code-removal branch? https://review.openstack.org/#/c/9696/ | 21:55 |
| vishy | jog0: cool. look in the sqlalchemy/api.py for syntax (iirc it is .options(joindload('hosts')) or some such | 21:58 |
| *** matwood has joined #openstack-dev | 22:01 | |
| *** jeblair has quit IRC | 22:02 | |
| jog0 | vishy: fixed, and retesting | 22:03 |
| *** nati has joined #openstack-dev | 22:03 | |
| *** ayoung has quit IRC | 22:04 | |
| *** armaan has left #openstack-dev | 22:04 | |
| *** markmcclain has quit IRC | 22:06 | |
| *** sniperd has quit IRC | 22:10 | |
| *** sniperd has joined #openstack-dev | 22:10 | |
| *** zul has quit IRC | 22:12 | |
| *** eglynn has quit IRC | 22:17 | |
| *** sacharya has quit IRC | 22:18 | |
| *** galstrom has quit IRC | 22:20 | |
| *** rafaduran has quit IRC | 22:20 | |
| *** andrewbogott_ has joined #openstack-dev | 22:22 | |
| *** andrewbogott_ has joined #openstack-dev | 22:22 | |
| *** andrewbogott has quit IRC | 22:22 | |
| *** andrewbogott_ is now known as andrewbogott | 22:22 | |
| BLZbubba | if i see references to "req-41b77e37-2b31-4392-a020-96de39580675" in the logs, where do I look to track that back to an actual request? | 22:23 |
| *** Gordonz has quit IRC | 22:25 | |
| bcwaldon | BLZbubba: you need to save the req id returned in the http headers | 22:25 |
| bcwaldon | BLZbubba: should be x-openstack-request-id | 22:26 |
| bcwaldon | BLZbubba: actually, x-compute-request-id from nova | 22:26 |
| bcwaldon | BLZbubba: we're moving towards x-openstack-request-id, afaik | 22:27 |
| BLZbubba | well i'm tracking down an unexplained reboot, and i got that req from the nova-compute logs | 22:27 |
| *** edygarcia has quit IRC | 22:27 | |
| BLZbubba | now i'm trying to figure out who/what generated that request | 22:28 |
| *** kbringard has quit IRC | 22:28 | |
| bcwaldon | BLZbubba: yeah, if its reproducible, then just make the request again and save that header | 22:28 |
| bcwaldon | BLZbubba: otherwise, you're stuck grepping the logs for that req-* | 22:28 |
| BLZbubba | not reproduceable, 2 machines just rebooted last night for no known reason | 22:29 |
| bcwaldon | BLZbubba: weird, so you're looking for a request that might have made that happen? | 22:29 |
| bcwaldon | BLZbubba: or you have a req id for the request in question | 22:29 |
| BLZbubba | i tried grepping but so far i have only seen it in nova-compute.log on the compute node, and in nova-network.log | 22:30 |
| *** ncode has quit IRC | 22:30 | |
| bcwaldon | BLZbubba: ok, not much I can do to help you then | 22:30 |
| bcwaldon | BLZbubba: just the info I already gave you :( | 22:30 |
| BLZbubba | right, i just have the AUDIT line from nova-compute and i want to know where those requests come from | 22:30 |
| BLZbubba | otherwise the audits aren't too useful :P | 22:30 |
| bcwaldon | BLZbubba: ah, I see | 22:30 |
| bcwaldon | BLZbubba: so those audit lines arent logged with a req id? | 22:31 |
| BLZbubba | yes there is a req id, but i don't see the req id in the api logs | 22:31 |
| bcwaldon | ah! | 22:31 |
| bcwaldon | BLZbubba: well that makes no sense | 22:31 |
| BLZbubba | AUDIT nova.compute.manager [req-41b77e37-2b31-4392-a020-96de39580675 None None] Rebooting instance 155dcf68-48c5 | 22:31 |
| *** nati has quit IRC | 22:31 | |
| BLZbubba | -4e8b-9680-35b05968e246 | 22:31 |
| BLZbubba | yeah i was just trying to figure out where i should expect to see the original request | 22:32 |
| bcwaldon | BLZbubba: maybe the context object is being recreated and generating a new req id | 22:32 |
| bcwaldon | BLZbubba: I would expect to see None and None be a project and user on the api server | 22:32 |
| BLZbubba | so nova-api is still the best bet? i can at least narrow it down by the time and maybe it will be obvious | 22:33 |
| *** alrs has quit IRC | 22:34 | |
| bcwaldon | BLZbubba: yes, I'm willing to be theres a request that correlates by time and function, but not req-id | 22:34 |
| *** danwent has quit IRC | 22:37 | |
| *** pixelbeat has joined #openstack-dev | 22:42 | |
| *** dachary has quit IRC | 22:44 | |
| *** dryan has joined #openstack-dev | 22:48 | |
| *** mattray has quit IRC | 22:48 | |
| *** gyee has joined #openstack-dev | 22:56 | |
| BLZbubba | hmm it looks like nova-api restarted about that time, but the machine didn't | 23:03 |
| *** mnewby has quit IRC | 23:04 | |
| *** e1mer has joined #openstack-dev | 23:05 | |
| *** e1mer has joined #openstack-dev | 23:05 | |
| *** arosen2 has quit IRC | 23:07 | |
| *** armaan has joined #openstack-dev | 23:08 | |
| *** belliott has quit IRC | 23:09 | |
| *** jog0 has quit IRC | 23:09 | |
| *** belliott has joined #openstack-dev | 23:09 | |
| *** jog0 has joined #openstack-dev | 23:09 | |
| *** torgomatic has quit IRC | 23:10 | |
| *** torgomatic has joined #openstack-dev | 23:10 | |
| *** andrewbogott has quit IRC | 23:10 | |
| *** bhuvan has quit IRC | 23:11 | |
| *** z has quit IRC | 23:11 | |
| *** bhuvan has joined #openstack-dev | 23:12 | |
| *** z has joined #openstack-dev | 23:12 | |
| *** z is now known as Guest53769 | 23:13 | |
| *** Mandell_ has joined #openstack-dev | 23:17 | |
| *** Mandell has quit IRC | 23:17 | |
| *** jeblair has joined #openstack-dev | 23:17 | |
| *** arosen1 has joined #openstack-dev | 23:22 | |
| *** everett-toews has quit IRC | 23:22 | |
| *** dtroyer is now known as dtroyer_zzz | 23:24 | |
| *** anniec has joined #openstack-dev | 23:31 | |
| *** armaan has left #openstack-dev | 23:31 | |
| *** anniec_ has joined #openstack-dev | 23:33 | |
| adam_g | mtaylor: ping | 23:33 |
| mtaylor | adam_g: pong | 23:34 |
| *** cmagina has quit IRC | 23:34 | |
| *** ncode has joined #openstack-dev | 23:35 | |
| *** anniec has quit IRC | 23:37 | |
| *** anniec_ is now known as anniec | 23:37 | |
| *** cmagina has joined #openstack-dev | 23:37 | |
| adam_g | mtaylor: hey, trying to adjust our build system to take into account the new versioning magic you've added. i'm looking at openstack/common/version.py now, do you see this changing at all anytime soon? http://paste.ubuntu.com/1088965/ | 23:37 |
| adam_g | i need to adjust the version string for the packages we're building to something dpkg will understand | 23:38 |
| mtaylor | adam_g: versioning scheme should not change | 23:39 |
| mtaylor | adam_g: the code itself I _just_ split from openstack/common into a standalone library | 23:39 |
| mtaylor | but that should not affect the thing you're talking about | 23:40 |
| mtaylor | adam_g: also - a couple of the client libs may need to the epochs added to their version on your side. sorry about that | 23:40 |
| adam_g | mtaylor: well, last week sdist produced something like client-0.1.1.tar.gz, this week its client-0.1.1.2.$hash.tar.gz | 23:41 |
| mtaylor | adam_g: it will produce both | 23:43 |
| mtaylor | depending on whether or not it's a real release or not | 23:43 |
| mtaylor | a client lib release of 0.1.1 will produce client-0.1.1.tar.gz | 23:43 |
| mtaylor | the next commit to git will cause us to generate client-0.1.1.1.$hash.tar.gz and upload that to tarballs.o.o but not pypi | 23:44 |
| mtaylor | the next will be client-0.1.1.2.$hash.tar.gz | 23:44 |
| mtaylor | until the next tagged release, which would be, well, whatever it's tagged as, but likely either client-0.1.2.tar.gz or client-0.2.0.tar.gz | 23:44 |
| mtaylor | yeah? | 23:44 |
| jog0 | vishy: I am having problems rying the joinedload option | 23:45 |
| jog0 | hosts is a function with a wrapper | 23:46 |
| adam_g | mtaylor: yeah, the scheme makes sense. i still need to do some tweaking on our side, though.. client-0.1.1.2.a814c15 and client-0.1.1.2.9e5eae1 aren't very helpful. its fine, though, if you dont expect the scheme to change | 23:50 |
| mtaylor | adam_g: I don't expect it to change | 23:50 |
| mtaylor | adam_g: btw - there will never be client-0.1.1.2.a814c15 AND client-0.1.1.2.9e5eae1 | 23:50 |
| mtaylor | adam_g: if there is, it's a MASSIVE systemic bug | 23:50 |
| adam_g | mtaylor: the 0.1.1.$foo.hash gets bumped with each commit? | 23:51 |
| mtaylor | the final number before the sha will always increase | 23:51 |
| mtaylor | yes | 23:51 |
| mtaylor | it's number of commits since last tag | 23:51 |
| adam_g | ok, super | 23:51 |
| mtaylor | yeah. the sha is there purely so that you can find the relevant commit back inthe repo - informational only | 23:51 |
| *** salv-orlando has joined #openstack-dev | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!