| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-12-13 | |||
| 07:35:57 | openstackgerrit | OpenStack Proposal Bot proposed openstack/nova master: Imported Translations from Zanata https://review.openstack.org/524795 | |
| 08:40:12 | openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: [libvirt] Add _set_vcpu_realtime_scheduler() https://review.openstack.org/527630 | |
| 08:40:13 | openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Add _set_vcpu_pinning() and _set_numa_memory() https://review.openstack.org/527631 | |
| 09:03:15 | openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Add _set_vcpu_pinning() and _set_numa_memory() https://review.openstack.org/527631 | |
| 09:17:47 | mdbooth | mnaser: melwitt Reading back scroll. Yes, a per-backend image cache was part of my plans both times I've submitted patches for a massive refactor of the imagebackend code. | |
| 09:18:38 | mdbooth | You'll hit a few architectural issues getting it landed, though. | |
| 09:18:52 | openstackgerrit | Yikun Jiang (Kero) proposed openstack/nova master: Add pagination and changes-since for instance-actions https://review.openstack.org/326326 | |
| 09:21:39 | mdbooth | Firstly, the image cache code is unnecessarily and messily embedded implicitly in each individual imagebackend already. All that code needs to be cleaned up. In itself that's not too hard, but the tests are also hugely problematic. Ultimately, fixing the tests to a point where it was possible to refactor the code sunk me twice. There are a huge number of tests in there, and in general the quality is rock bottom. | |
| 09:23:19 | mdbooth | However, if you get to the point where you've unpicked the imagecache from the imagebackend code, you still have to handle both layouts. | |
| 09:23:58 | mdbooth | Currently, the only source of truth for the layout of an instance is the config of the compute host is resides on. | |
| 09:24:54 | mdbooth | So if we're to have a transitional period where we're altering the on-disk layout, which is what changing the image cache would amount to, we need to have a per-instance datastore for driver-specific data. | |
| 09:25:22 | mdbooth | Incidentally, my plan was to add this to the BDM so it could actually be per-disk. | |
| 09:26:43 | mdbooth | But if you can get there, yes the current design of the imagecache makes no sense | |
| 09:27:04 | mdbooth | Well, no longer makes any sense | |
| 09:27:27 | mdbooth | It should have been refactored long ago the first time we implemented an alternate backend | |
| 09:36:53 | openstackgerrit | Yikun Jiang (Kero) proposed openstack/nova master: Add pagination and changes-since for instance-actions https://review.openstack.org/326326 | |
| 09:41:06 | kashyap | mdbooth: You might want to spell that documentation somewhere to whoever that is planning to go that route... | |
| 09:42:01 | kashyap | Maybe you did in all those previous iterations. In that case, disregard me :-) | |
| 09:42:14 | mdbooth | kashyap: I wrote the code :) | |
| 09:42:21 | mdbooth | Then I threw it away. | |
| 09:42:25 | mdbooth | The I wrote it again! | |
| 09:42:29 | kashyap | mdbooth: Whoops | |
| 09:42:33 | mdbooth | Then I threw that away, too. | |
| 09:42:44 | kashyap | Hence the "sunk me twice", now that nice expression makes sense | |
| 09:42:59 | kashyap | s/makes sense/makes doubly sense/ | |
| 09:43:01 | mdbooth | Each one took me a couple of months. | |
| 09:43:23 | mdbooth | Throwing away 2 months of effort is good for the soul. | |
| 09:43:28 | kashyap | :-) | |
| 09:44:00 | kashyap | mdbooth: Just out of curiosity, as you meditated on this problem space, how badly does this affect us long-term? | |
| 09:45:04 | mdbooth | kashyap: The main issue in imagebackend, and the reason it needs a refactor like either of the ones I already did, is that every time somebody comes along and proposes a design change it's basically infeasible. | |
| 09:45:24 | mdbooth | Right now we have a big tangled ball of string that works. | |
| 09:45:46 | mdbooth | That's great, but you can't easily make it do anything else without untangling it first. | |
| 09:46:34 | mdbooth | So, a new backend which works substantially differently to the existing ones. | |
| 09:46:43 | kashyap | Hmm | |
| 09:46:51 | mdbooth | Changing the on-disk layout | |
| 09:47:00 | gibi | dansmith: I hit https://review.openstack.org/#/c/517119/ but as I see Jay will respin it one more time so I will hit it with a +2 after that | |
| 09:47:07 | mdbooth | e.g. because you want a more efficient imagecache | |
| 09:47:14 | mdbooth | or because you want to use libvirt storage pools | |
| 09:48:02 | kashyap | mdbooth: Changing the on-disk layout to what? | |
| 09:48:21 | kashyap | mdbooth: If I'm asking uproductive questions, ignore me. I'll do some looking up :-) | |
| 09:48:26 | mdbooth | Well the 2 proposals above have both actually been made. | |
| 09:48:43 | kashyap | I know you've written to the list and on changes you proposed with your thoughts on this | |
| 09:48:47 | openstackgerrit | Lee Yarwood proposed openstack/nova master: conf: Do not inherit image signature props with snapshots https://review.openstack.org/527046 | |
| 09:48:48 | mdbooth | Changing the image cache is an on-disk change, because backing images are now in a different place. | |
| 09:49:00 | kashyap | Ah, right. | |
| 09:49:05 | mdbooth | The existing code assumes all sorts of things in obtuse ways. | |
| 09:49:22 | kashyap | Will this be brought up at the upcoming PTG? | |
| 09:49:43 | kashyap | Assuming you /others still have the appetite for it... :-) | |
| 09:49:49 | mdbooth | Using libvirt storage pools requires an on-disk change, because we don't layout our disks in a way that's easy to manage with storage pools. | |
| 09:50:21 | mdbooth | It's not specifically on my agenda right now. I rehash it from time to time (like this morning) when somebody else brings it up. | |
| 09:50:44 | mdbooth | I've done the work twice already. | |
| 09:53:30 | kashyap | Nod; thanks for the explanation. | |
| 10:03:12 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: Add PCI NUMA policy fields https://review.openstack.org/527470 | |
| 10:03:13 | openstackgerrit | Stephen Finucane proposed openstack/nova master: trivial: Pass InstanceNUMATopology to consume_request https://review.openstack.org/527471 | |
| 10:03:14 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Add PCI NUMA policies https://review.openstack.org/527472 | |
| 10:03:14 | openstackgerrit | Stephen Finucane proposed openstack/nova master: trivial: Modify signature of _filter_non_requested_pfs https://review.openstack.org/527473 | |
| 10:12:29 | openstackgerrit | Ya Lian Pan proposed openstack/nova master: z/VM Driver: Spawn and destroy function of z/VM driver https://review.openstack.org/527658 | |
| 10:13:42 | openstackgerrit | Chason Chan proposed openstack/nova master: Fix the bug report link of API Guide https://review.openstack.org/527660 | |
| 10:32:46 | openstackgerrit | Stephen Finucane proposed openstack/nova master: zuul: Move legacy jobs to project https://review.openstack.org/514309 | |
| 10:32:47 | openstackgerrit | Stephen Finucane proposed openstack/nova master: zuul: Remove WSGI functional tests https://review.openstack.org/527140 | |
| 11:38:05 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: remove pagesize from __init__ of InstanceNUMATopology https://review.openstack.org/485553 | |
| 11:38:05 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: remove related pinning from __init__ of InstanceNUMATopology https://review.openstack.org/485554 | |
| 11:38:06 | openstackgerrit | Stephen Finucane proposed openstack/nova master: objects: remove cpuset_reserved from __init__ of InstanceNUMATopology https://review.openstack.org/466030 | |
| 11:39:17 | stephenfin | dansmith: When you're about, could you take a punt at https://review.openstack.org/#/q/topic:bug/1636338 Guess you're best placed for them | |
| 11:39:40 | stephenfin | dansmith: Or stick em on your list for the new year. Either-or (they're just there a looong time) | |
| 11:43:10 | openstackgerrit | Li Xipeng proposed openstack/nova master: Fix bug case by none token context https://review.openstack.org/522112 | |
| 12:10:27 | openstackgerrit | Lajos Katona proposed openstack/nova master: Extend ServerMovingTests with custom resources https://review.openstack.org/497399 | |
| 12:55:05 | openstackgerrit | Chris Dent proposed openstack/nova master: [placement] Enable limiting GET /allocation_candidates https://review.openstack.org/513526 | |
| 12:58:41 | gmann | nova api meeting in 2 min.. | |
| 13:41:12 | openstackgerrit | Merged openstack/nova master: Re-use existing ComputeNode on ironic rebalance https://review.openstack.org/508555 | |
| 13:46:47 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Transform instance.exists notification https://review.openstack.org/403660 | |
| 14:01:27 | openstackgerrit | Chen Hanxiao proposed openstack/nova master: libvirt: guest: introduce blockStats instead of domain.blockStats https://review.openstack.org/526833 | |
| 14:02:04 | cdent | edleafe: are you actually asking for changes on that commit message or just pointing out some potential changes if there's another version. On the random stuff I'm not sure that your version is more clear, at least not to me. | |
| 14:02:39 | edleafe | cdent: Well, you *did* push another version :) | |
| 14:03:10 | cdent | yeah, and totally forgot the commit message stuff because I was focused on the merge conflict (and forgot about your suggestions because no vote) | |
| 14:03:28 | edleafe | cdent: The random comment simply seemed to describe the implementation rather than the effect of changing the settings | |
| 14:04:32 | cdent | right, the commit message is for describing the implementation, the conf help and the reno are for describing why/how someone might use the conf setting | |
| 14:10:05 | mriedem | gibi: replied to your questions in https://review.openstack.org/#/c/507473/ | |
| 14:13:23 | openstackgerrit | Kashyap Chamarthy proposed openstack/nova master: conf: libvirt: Cleanup CPU modelling related options https://review.openstack.org/527691 | |
| 14:16:44 | stephenfin | kashyap: Per ^, you'll probably like this https://review.openstack.org/#/c/526012/ | |
| 14:21:17 | kashyap | stephenfin: OCD: If you have to respin, please captialize the proper noun: s/nova/Nova/ :-) | |
| 14:21:44 | stephenfin | kashyap: Ahem https://docs.openstack.org/doc-contrib-guide/writing-style/openstack-components.html | |
| 14:21:53 | stephenfin | My OCD trumps yours | |
| 14:22:14 | kashyap | Haha | |
| 14:22:18 | gibi | mriedem: thanks for the answers, I'm +2 now | |
| 14:22:46 | kashyap | Okay, my brain just feels a twitch when a sentence starts with a smaller case letter. | |
| 14:22:54 | mriedem | gibi: cool. i'm going to update the nits in the last change in the series too just to get that going | |
| 14:23:17 | mnaser | mdbooth: thanks for the informative text. The more I look into this the more I see that it’s too much for me to bite to make that image backend change | |
| 14:24:18 | mdbooth | mnaser: Hopefully that's not true. I'd just say don't underestimate the size of the chunk :) | |
| 14:24:33 | mdbooth | And I completely agree with your initial conclusion. | |
| 14:25:10 | mnaser | mdbooth: the thing is we’re fairly pressed on time and a change like this would probably take a long time to get through | |
| 14:25:15 | kashyap | stephenfin: Nice Olso change in that; Zuul seems to complain for tox-py35, must've noticed it | |
| 14:25:19 | mdbooth | Oh, right. | |
| 14:25:29 | mdbooth | Yeah, this isn't going to be a quick fix. | |
| 14:25:43 | stephenfin | kashyap: I did indeed. I'll be fixing it shortly | |
| 14:25:45 | gibi | mriedem: hit me with the link when you are done and we can push the last patch through as well | |
| 14:26:09 | mnaser | Figuring out how to work with Glance and multiple locations seems far more feasable. I think. | |
| 14:26:10 | mdbooth | mnaser: Although, the current state of that code is the sum of a bunch of quick fixes... | |
| 14:26:13 | kashyap | stephenfin: Thanks for the fast +2 on that change! | |
| 14:26:17 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Add sample test for instance audit https://review.openstack.org/480955 | |