Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-12-04
14:45:22 sean-k-mooney2 gibi: ah yes that is a good point. i would prefer to use nested resouce providers as logically the mdev resouce are a nested resouce of a parent device. in my case a pcie device but technically mdevs could be childern of device on buses other the pci bus.
14:46:39 gibi sean-k-mooney2: I agree that the best would be to have nest RPs for it
14:46:52 efried jaypipes bauzas alex_xu Oh, I wasn't following the conversation in here; I left a (verbose, as usual) comment.
14:47:05 jaypipes efried: on what?
14:48:15 jaypipes efried: in any case, I'm currently adding func tests that add member_of and uuid filters to a tree filter.
15:01:03 alex_xu efried: I don't the tree and resources will work as your expected. The 'GET /resource_providers' just a simple filtering, it is different with the filtering of 'GET /allocation_candidates'
15:02:40 efried jaypipes alex_xu Sorry, I was talking about https://review.openstack.org/#/c/377215/
15:02:44 efried I'm coming at it from the perspective of how I think the REST API should work, not looking at the code.
15:03:14 efried The func tests jaypipes is suggesting should bear out what I said in my comment; and if not, it's a bug that we can fix, right?
15:04:43 openstackgerrit Takashi NATSUME proposed openstack/nova master: [placement] Add sending global request ID in delete https://review.openstack.org/525231
15:07:13 jianghuaw_ bauzas, the vGPU relative calls have been there since the minimal version of XS supported by nova. So there is no need to check XenServer's version.
15:07:38 openstackgerrit Jay Pipes proposed openstack/nova master: placement: adds REST API for nested providers https://review.openstack.org/384807
15:07:38 stephenfin dansmith: Question on this patch for you https://review.openstack.org/#/c/514309/
15:07:38 openstackgerrit Jay Pipes proposed openstack/nova master: placement: allow filter providers in tree https://review.openstack.org/377215
15:07:39 openstackgerrit Jay Pipes proposed openstack/nova master: placement: update client to set parent provider https://review.openstack.org/385693
15:07:52 jaypipes bauzas, alex_xu, efried: k, func tests added for tree filter + uuid/member_of
15:08:19 dansmith stephenfin: where?
15:08:29 bauzas jianghuaw_: gtk
15:08:36 stephenfin dansmith: The 'tempest-dsvm-neutron-src-oslo.versionedobjects' job
15:08:53 stephenfin dansmith: I migrated that because it seemed we were the only caller
15:09:15 bauzas jianghuaw_: kaboom then
15:09:34 stephenfin But I see now that there are many 'tempest-dsvm-neutron-src-{library}' jobs and '{library}' is usually wildcarded, so that should probably live in oslo.versionedobjects
15:09:42 jianghuaw_ bauzas, thanks:-)
15:10:26 dansmith stephenfin: sorry, I blame the early hour for my confusion but.. what are you asking? if other projects have o.vo jobs?
15:11:21 stephenfin dansmith: Yes. Just a sanity check
15:12:18 dansmith stephenfin: tbh, that's not a thing I know and I would expect it to be more of a core oslo thing, policy for running library jobs against projects that use them and such
15:12:42 dansmith stephenfin: I think nova, neutron, ironic at least use o.vo
15:12:49 jianghuaw_ bauzas, I guess this one would be much easier as you ever +2/+w on it:-) https://review.openstack.org/#/c/516217
15:13:11 bauzas jianghuaw_: next one :)
15:13:25 jianghuaw_ cool:-)
15:13:26 efried jaypipes It would be nice to have a func test for resources + tree -- I think that was a main concern of alex_xu's. But we can do it in a followup too.
15:13:27 stephenfin dansmith: That's fair. I picked on you as you were the one that enabled the test for nova (commit ac96f95 in openstack-infra/project-config). No good deed :)
15:13:34 stephenfin I'll go bug AJaeger
15:13:50 dansmith stephenfin: probably at the behest (and handholding) of some infra deity :)
15:17:27 openstackgerrit Takashi NATSUME proposed openstack/nova master: Fix 500 error while passing 4-byte unicode data https://review.openstack.org/407514
15:20:53 efried jaypipes I was about to rebase "make _Provider private" -- did you have any feedback on uuid_set() (https://review.openstack.org/#/c/520243/) so I can hit it at the same time?
15:37:22 openstackgerrit Eric Fried proposed openstack/nova master: ProviderTree.uuid_set() https://review.openstack.org/520243
15:37:22 openstackgerrit Eric Fried proposed openstack/nova master: Make _Provider really private https://review.openstack.org/523932
15:37:28 efried jaypipes ^
15:40:00 mdbooth dansmith: You mentioned last week that you didn't think I could add a unique constraint to the uuid column in the first instance. You also say the same explicitly in a comment, so I assume there's a good reason. Do you recall what it is?
15:41:31 dansmith mdbooth: you can if it's nullable I think, but I think I thought you were asserting that you were going to create the column and make it non-nullable/unique right away and that's what I was trying to say was not okay
15:41:57 mdbooth dansmith: Yeah, I was going for unique/nullable.
15:42:17 mdbooth Any reason I shouldn't add the unique constraint to your patch?
15:42:49 dansmith if you can add it to the schema without code changes or moving data, then it's cool to add it yeah
15:43:05 mdbooth Will do, thanks
15:45:10 tssurya mriedem: ping
15:51:52 openstackgerrit Balazs Gibizer proposed openstack/nova master: Add late server group policy check to rebuild https://review.openstack.org/525242
15:51:55 mdbooth dansmith: Second question: update by uuid. I'm assuming that when you were adding this first time round you anticipated needing this. I don't think I need it. I could add it for completeness, or omit it until there's a caller. Personal preference would be the latter.
15:53:37 dansmith mdbooth: I probably had it for tests or something, but sure, no need to have it if there is no user
15:59:22 openstackgerrit Jackie Truong proposed openstack/python-novaclient master: Microversion 2.57 - Add trusted_image_certificates https://review.openstack.org/500396
16:07:37 mdbooth dansmith: Thanks
16:13:09 ildikov mriedem: hi, I updated the patch with the fixes to your comments
16:13:30 ildikov mriedem: I didn't have the stomach as of yet at least to do the split...
16:13:45 mriedem ildikov: ok
16:14:30 ildikov mriedem: if you have time plz continue to review and I will fix it in either one or two patches
16:15:16 mriedem i hope to get to it today, but have to work on some other stuff this morning
16:16:45 ildikov appreciated if you can get to it today
16:33:24 tssurya mriedem: I am stuck on this - https://bugs.launchpad.net/nova/+bug/1729806/comments/1 , because both those columns are of type "Text()" and SQL doesn't allow text type columns to have a unique constraint; any pointers on how you usually do this in nova ?
16:33:26 openstack Launchpad bug 1729806 in OpenStack Compute (nova) "nova-manage update_cell does not check if the same combination of transport_url and database_connection already exists" [Low,In progress] - Assigned to Surya Seetharaman (tssurya)
16:33:37 mriedem ildikov: i looked at the diff, comments in https://review.openstack.org/#/c/330285/
16:34:50 openstackgerrit Eric Fried proposed openstack/nova master: Make _Provider really private https://review.openstack.org/523932
16:36:24 dansmith tssurya: you could check before create, which isn't as good as a constraint, but it'd be an easy thing to do
16:36:29 openstackgerrit Silvan Kaiser proposed openstack/nova master: Added mount fstype based validation of Quobyte mounts https://review.openstack.org/522245
16:36:32 ildikov mriedem: tnx!
16:37:09 mriedem ildikov: and another one just now
16:37:18 mriedem ildikov: the shelved offloaded case probably needs some thought
16:37:23 mriedem could also see what johnthetubaguy thinks
16:37:39 ildikov mriedem: ok, I'll wait a bit longer before touching it :)
16:37:47 mriedem for all intents and purposes, a reserved volume is going to be the same as an in-use volume to nova
16:37:51 ildikov mriedem: sigh, I will look into the comments shortly
16:38:16 johnthetubaguy I was a touch worried about the change in shelved offloaded
16:38:28 mriedem there is an api behavior change with attaching a volume to a shelved offloaded instance with the new flow,
16:38:37 mriedem where the volume is left 'reserved' rather than 'in-use'
16:38:42 johnthetubaguy yeah, that seemed nasty
16:38:44 mriedem i could see client side app tooling failing / timing out on that
16:38:46 mriedem like tempest even
16:38:57 mriedem lots of client code waits for an attachment to go to in-use before it considers it done
16:39:07 johnthetubaguy yes, good point
16:39:10 tssurya dansmith: Yes, that is what I am doing in my review, I was just trying to add the constraint part too
16:39:42 dansmith tssurya: sounds like we might not be able to do the constraint part.. since urls can't be large, I'm not sure converting it to something other than text is a good idea
16:40:07 johnthetubaguy mriedem: I did raise that in a previous version, but forgot about it shortly after, +1 on it being bad
16:40:09 dansmith tssurya: we actually have some long urls ourselves, since we encode all the HA hosts in the transport url, and I think we do the same for our galera endpoints
16:40:40 efried cdent dansmith Did you see my "themed CI results" last week? I made one monochrome for cdent (http://184.172.12.213/manual/htmlify_logs/logs/n-cpu.txt.gz?theme=monochrome) and one solarized for both of you (http://184.172.12.213/manual/htmlify_logs/logs/n-cpu.txt.gz?theme=solarized)
16:40:55 tssurya dansmith: oh okay, yea then I will not pursue it, this came up https://review.openstack.org/#/c/523817/ in parallel to my patchset, so I thought I would investigate this
16:41:06 dansmith tssurya: cool
16:41:09 tssurya dansmith: thanks then, I will remove the constraint part then
16:41:12 mriedem edleafe: i'm going to make a concerted push on the alternate hosts series today, do you plan on rebasing https://review.openstack.org/#/q/topic:bp/return-alternate-hosts+status:open to resolve the merge conflicts at the top?
16:41:43 dansmith efried: I did not, but cool.. that still ends up with ascii in the logs if I curl download it though right?
16:42:25 AJaeger_ sdague: could you review https://review.openstack.org/#/c/524353/ ,please? This updates jobs for os-api-ref and gets rid of tools/tox_install.sh
16:42:25 efried dansmith I've actually got solutions that go both ways.
16:42:30 jaypipes efried: you are currently rebasing the series that ends in 520246, yes?
16:42:32 dansmith efried: ....
16:42:40 edleafe mriedem: I was waiting until there was renewed interest in the series. I can start rebasing now.
16:42:48 dansmith efried: well, from the purely browser-based view, solarized looks good to my eyes :)
16:42:52 mriedem edleafe: yup, it's my #1 priority today
16:43:01 efried dansmith I think what I've got on that server right now preserves the codes, because that's the way I prefer it, and that's my server :)
16:43:02 mriedem edleafe: i plan on having that series +2ed by thursday
16:43:32 efried dansmith Okay, duly noted. The way I've got it set up, it's fairly easy to add themes, and (in case you didn't notice) there are new links at the top allowing you to flip between themes.
16:43:32 edleafe mriedem: on it
16:43:52 dansmith efried: yeah, curl gives me ansified files

Earlier   Later