| *** mriedem has quit IRC | 00:02 | |
| *** efried has joined #openstack-placement | 00:05 | |
| *** tetsuro has joined #openstack-placement | 00:17 | |
| alanmeadows | mnaser: hmm, so running this in an environment | 00:27 |
|---|---|---|
| alanmeadows | with clearly missing allocations entries | 00:27 |
| alanmeadows | I don't see any --set advisories | 00:27 |
| alanmeadows | a whole lot of orphan cleanup though which is not what I expected | 00:30 |
| alanmeadows | http://paste.openstack.org/show/743139/ | 00:31 |
| alanmeadows | might of been my bad for miss-reading | 00:35 |
| alanmeadows | I scanned it and assumed this would handle creating missing allocations (not just fixing them) | 00:35 |
| *** tetsuro has quit IRC | 01:27 | |
| *** tetsuro has joined #openstack-placement | 01:56 | |
| *** tetsuro has quit IRC | 02:03 | |
| *** tetsuro has joined #openstack-placement | 03:13 | |
| *** openstackgerrit has joined #openstack-placement | 03:17 | |
| openstackgerrit | Tetsuro Nakamura proposed openstack/placement master: Add upgrade status check for missing root ids https://review.openstack.org/632599 | 03:17 |
| openstackgerrit | Tetsuro Nakamura proposed openstack/placement master: Use local config for placement-status CLI https://review.openstack.org/632600 | 03:17 |
| openstackgerrit | Dakshina Ilangovan proposed openstack/nova-specs master: Nova changes to support Local Resource Management that uses Resource Management Daemon https://review.openstack.org/630817 | 06:44 |
| *** e0ne has joined #openstack-placement | 07:57 | |
| *** tssurya has joined #openstack-placement | 08:17 | |
| *** e0ne has quit IRC | 08:33 | |
| *** helenafm has joined #openstack-placement | 08:40 | |
| *** tetsuro has quit IRC | 09:11 | |
| *** ttsiouts has joined #openstack-placement | 09:26 | |
| *** ttsiouts has quit IRC | 09:47 | |
| *** ttsiouts has joined #openstack-placement | 09:48 | |
| *** ttsiouts_ has joined #openstack-placement | 09:49 | |
| *** ttsiouts has quit IRC | 09:50 | |
| *** cdent has joined #openstack-placement | 10:45 | |
| cdent | just a heads up that the release of os-resource-classes has broken the placement functional tests. I'm fixing it | 10:47 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Update standard resource class counts in tests https://review.openstack.org/632696 | 11:01 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Increase loop size on _ensure_aggregate https://review.openstack.org/632697 | 11:01 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Adjust database connection pool config in perfload tests https://review.openstack.org/632698 | 11:01 |
| *** ttsiouts_ has quit IRC | 11:01 | |
| *** ttsiouts has joined #openstack-placement | 11:02 | |
| cdent | gibi, efried, jaypipes : the bottom one in that ^ stack is important, the rest are simply nice to have | 11:02 |
| *** ttsiouts has quit IRC | 11:06 | |
| *** e0ne has joined #openstack-placement | 11:26 | |
| *** ttsiouts has joined #openstack-placement | 11:28 | |
| *** cdent has quit IRC | 12:34 | |
| *** ttsiouts has quit IRC | 12:53 | |
| *** cdent has joined #openstack-placement | 12:54 | |
| *** ttsiouts has joined #openstack-placement | 12:54 | |
| *** ttsiouts has quit IRC | 12:58 | |
| *** ttsiouts has joined #openstack-placement | 12:59 | |
| jaypipes | cdent: ack | 13:07 |
| cdent | that reminds me, forgot to leave a comment on that | 13:07 |
| cdent | hmm, it's also all broken | 13:08 |
| jaypipes | cdent: see my response. | 13:15 |
| jaypipes | cdent: generally, I believe we should be pinning all dependent libraries. | 13:15 |
| cdent | I tend to believe the opposite :) | 13:16 |
| cdent | I think deployments should pin | 13:16 |
| cdent | but I think development should break asap | 13:16 |
| jaypipes | cdent: the problem with deployments pinning is that it puts the onus on deployers to know all the specific dependencies of all the interrelated software bits in their application. But it's the developers of the application(s) themselves that know the dependencies best, IME. | 13:17 |
| cdent | which I think is something to do with why lower-constraints exists, but... | 13:17 |
| jaypipes | cdent: also, please /see my general disdain for all things deployment ;) | 13:17 |
| cdent | same | 13:18 |
| jaypipes | :) | 13:18 |
| cdent | imma leave it for now because if we change it for os-resource-classes we probably should for os-traits as well (more details when I respond on the review) | 13:25 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Update standard resource class counts in tests https://review.openstack.org/632696 | 13:26 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Increase loop size on _ensure_aggregate https://review.openstack.org/632697 | 13:27 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Adjust database connection pool config in perfload tests https://review.openstack.org/632698 | 13:27 |
| gibi | cdent: I'm +2 on the os-resource-classes version bump | 13:52 |
| jaypipes | gibi: what are your thoughts about pinning os-resource-classes (and os-traits) in nova? | 13:53 |
| gibi | cdent: on the second in that stack I have a comment in the commit message https://review.openstack.org/#/c/632697/2//COMMIT_MSG | 13:55 |
| cdent | gibi: right, thanks, clearly I got ahead of myself | 13:55 |
| *** mriedem has joined #openstack-placement | 13:56 | |
| openstackgerrit | Chris Dent proposed openstack/placement master: Increase loop size on _ensure_aggregate https://review.openstack.org/632697 | 13:56 |
| cdent | fxied | 13:56 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Adjust database connection pool config in perfload tests https://review.openstack.org/632698 | 13:56 |
| gibi | jaypipes: I'm siding with cdent about pinning. In dev CI I'd like to see the breaking change in the lib as soon as possible. Especially for a library that mostly safe to no to pin as the change in os-traits and os-resource-classes are mostly additions | 13:56 |
| gibi | jaypipes: do we even pin libvirt and qemu for nova? | 13:57 |
| gibi | cdent: thanks, +2 | 14:01 |
| cdent | thank you | 14:01 |
| jaypipes | gibi: we don't pin anything in nova, AFAIK... | 14:01 |
| jaypipes | gibi: at least, none of the dep libs in requirements.txt have a hard pin (they are all >= constraints) | 14:02 |
| gibi | jaypipes: so do you suggest to start pinning our deps in nova, and start with os-resource-classes and os-traits | 14:03 |
| gibi | ? | 14:03 |
| gibi | jaypipes: in the past there was upper constraints somewhere, doesn't it? | 14:04 |
| gibi | jaypipes: we still have them https://github.com/openstack/requirements/blob/master/upper-constraints.txt | 14:04 |
| gibi | so basically we have both lower and upper | 14:04 |
| jaypipes | gibi: no, I'm not actually suggesting anything :) I'm just throwing out my thoughts. I personally think that all dependent libraries should be pinned at a known good state and that the build artifact for an application should be a binary blob of very specific known-to-work-with-each-other code. | 14:05 |
| jaypipes | gibi: it's just that we're working in Python here and not producing a single build artifact, unfortunately. | 14:06 |
| jaypipes | gibi: so I do understand that this way works for Python software applications like Nova. just that I prefer a more strict build artifact. | 14:06 |
| jaypipes | (though I freely admit I'm no packaging or release expert) | 14:07 |
| * cdent can feel jaypipes feeling around in the go static binaries, give me a tiny deployment, perhaps in an immutable, distroless container styleeee | 14:08 | |
| cdent | it would be nice | 14:08 |
| jaypipes | cdent: yep. zactly. | 14:09 |
| jaypipes | cdent: btw, I +W'd the bottom patch on that series. | 14:09 |
| cdent | thanks | 14:10 |
| *** mriedem has quit IRC | 14:10 | |
| cdent | edleafe: what does "Do you also have the "after" results for the perfload tests?" mean? | 14:36 |
| edleafe | cdent: One without the "distracting noise" that the patch is supposed to remedy | 14:37 |
| cdent | it will show up when that patch's tests run? | 14:37 |
| edleafe | I just wanted to be able to see both so that I could interpret better | 14:38 |
| edleafe | Right now it looks like a mutant DNA strand to me :) | 14:38 |
| edleafe | NVM, http://logs.openstack.org/98/632698/1/check/placement-perfload/67e0f93/logs/placement-perf.txt | 14:39 |
| cdent | it remains a mutant strand of dna | 14:40 |
| cdent | but without the 500s in it (which are in the results linked in the commit message) | 14:40 |
| * cdent takes a walk, biab | 14:41 | |
| *** cdent has quit IRC | 14:41 | |
| *** efried has quit IRC | 15:08 | |
| *** ttsiouts has quit IRC | 15:13 | |
| *** cdent has joined #openstack-placement | 15:14 | |
| *** ttsiouts has joined #openstack-placement | 15:14 | |
| *** ttsiouts has quit IRC | 15:19 | |
| *** e0ne has quit IRC | 15:30 | |
| *** efried has joined #openstack-placement | 15:30 | |
| *** e0ne has joined #openstack-placement | 15:33 | |
| *** e0ne has quit IRC | 15:33 | |
| *** e0ne has joined #openstack-placement | 15:33 | |
| *** ttsiouts has joined #openstack-placement | 15:35 | |
| gibi | jaypipes: thanks for the explanation. I see what you are after. | 15:46 |
| gibi | jaypipes: so I'm not against it, but it might need a wider communication and collaboration to make it happen what you imagine | 15:47 |
| gibi | jaypipes: for example, if you pin then we need to have a process to sometimes dump the pinned version | 15:48 |
| *** helenafm has quit IRC | 15:54 | |
| *** efried has quit IRC | 16:03 | |
| jaypipes | gibi: right, and I'm not recommending we change the whole way that dependencies are managed for nova :) just saying I personally prefer a different approach. | 16:25 |
| openstackgerrit | Merged openstack/placement master: Update standard resource class counts in tests https://review.openstack.org/632696 | 16:27 |
| gibi | jaypipes: understood :) | 16:29 |
| *** avolkov has quit IRC | 16:39 | |
| *** e0ne has quit IRC | 16:55 | |
| *** tssurya has quit IRC | 17:00 | |
| *** ttsiouts has quit IRC | 17:01 | |
| *** ttsiouts has joined #openstack-placement | 17:02 | |
| *** ttsiouts has quit IRC | 17:07 | |
| openstackgerrit | Merged openstack/placement master: Increase loop size on _ensure_aggregate https://review.openstack.org/632697 | 18:54 |
| openstackgerrit | Merged openstack/placement master: Adjust database connection pool config in perfload tests https://review.openstack.org/632698 | 18:54 |
| *** e0ne has joined #openstack-placement | 19:10 | |
| *** e0ne has quit IRC | 19:32 | |
| *** e0ne has joined #openstack-placement | 19:40 | |
| *** e0ne has quit IRC | 20:18 | |
| *** mriedem has joined #openstack-placement | 20:56 | |
| *** cdent has quit IRC | 22:39 | |
| *** mriedem has quit IRC | 23:48 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!