| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-01-26 | |||
| 18:18:55 | melwitt | johnthetubaguy: so on the race prevention thing, you were just thinking of checking the limit twice? once before object create and once after? | |
| 18:19:09 | johnthetubaguy | melwitt: yeah, I mean thats racey, but in a quota goes low sense | |
| 18:19:19 | melwitt | I guess you'd have to redo the count too | |
| 18:19:21 | melwitt | yeah | |
| 18:20:34 | johnthetubaguy | melwitt: yeah | |
| 18:20:57 | johnthetubaguy | melwitt: its just easier to reason about the correctness | |
| 18:21:17 | melwitt | yeah | |
| 18:22:43 | melwitt | that's actually a painfully easy fix for a bug we have open where someone was able to create 100+ more keypairs than the quota | |
| 18:23:24 | melwitt | well, I guess there's still a window there so I dunno how much it would lessen it | |
| 18:24:43 | andreykurilin | ok, I'm near PC | |
| 18:24:58 | johnthetubaguy | melwitt: I think it would mean we would delete too many though, I could be wrong | |
| 18:25:49 | melwitt | cool, thanks for all the comments on the review | |
| 18:26:33 | andreykurilin | johnthetubaguy SecretHelper is used at L851 - https://review.openstack.org/#/c/424256/10/novaclient/shell.py . It check args and if password is not there it create "prompt" | |
| 18:26:54 | andreykurilin | let's do it in Pike... | |
| 18:27:00 | bauzas | andreykurilin: I need to bail out, but I leave you do the change and ping one of us for single-approving the reno change | |
| 18:27:10 | openstackgerrit | Sergey Nikitin proposed openstack/nova: Change parent of diagnostics objects to NovaObject class https://review.openstack.org/394480 | |
| 18:27:10 | openstackgerrit | Sergey Nikitin proposed openstack/nova: Standardization of VM diagnostics info API. https://review.openstack.org/355540 | |
| 18:27:11 | andreykurilin | bauzas: I need a minute | |
| 18:27:12 | openstackgerrit | Sergey Nikitin proposed openstack/nova: Completed implementation of instance diagnostics for Xen https://review.openstack.org/399613 | |
| 18:27:12 | bauzas | andreykurilin: just tell the other core that I'm +2 | |
| 18:27:27 | bauzas | andreykurilin: we need the patch to be merged because we can tag the release | |
| 18:27:31 | bauzas | before* | |
| 18:27:42 | bauzas | andreykurilin: as I need the SHA1 of the merge commit | |
| 18:27:48 | snikitin | sfinucan, johnthetubaguy: Hi guys! I fixed your nits in diagnoistics. Could you please look? https://review.openstack.org/#/c/394480/ | |
| 18:28:04 | cdent | bauzas: i'm around for the next three hours, anything in partivular i should watching? | |
| 18:28:24 | bauzas | cdent: I think we addressed most of the concerns | |
| 18:28:31 | openstackgerrit | Andrey Kurilin proposed openstack/python-novaclient: Add release not for fixing token auth method https://review.openstack.org/425822 | |
| 18:28:38 | bauzas | cdent: the scheduler series is in good path, we're waiting for CI results | |
| 18:28:48 | andreykurilin | bauzas: you can add Depends-On | |
| 18:28:49 | cdent | awesome | |
| 18:29:07 | andreykurilin | bauzas: https://review.openstack.org/#/c/425822/1 | |
| 18:29:09 | bauzas | cdent: dansmith had a good valid point about how we should cache the min version | |
| 18:31:27 | bauzas | andreykurilin: fast-approved | |
| 18:31:53 | andreykurilin | bauzas: thanks | |
| 18:32:12 | bauzas | andreykurilin: I can't add a Depends-On as I wouldn't know the SHA1 | |
| 18:32:35 | bauzas | that's not a change-id that I need to provide | |
| 18:32:55 | andreykurilin | bauzas: why? you can use SHA1 from my commit and do not wait for merge-commit. I think it worked before | |
| 18:33:15 | bauzas | andreykurilin: because I'm not sure we won't get a merge commit | |
| 18:33:20 | andreykurilin | heh | |
| 18:33:48 | andreykurilin | ok, np. It looks like requirements team is ok to wait a bit:) | |
| 18:34:00 | bauzas | if I'm tagging based on your patch SHA1, then it would cut the release based on your local branch, not gerrit's | |
| 18:38:24 | bauzas | andreykurilin: we now support 2.41 ? | |
| 18:38:46 | bauzas | just writing release highlights | |
| 18:39:34 | andreykurilin | bauzas: Human beings have dreams. Even dogs have dreams, but not gerrit, gerrit is just a program. An imitation of life. Can a program write a symphony? Can a program turn a... canvas into a beautiful masterpiece?... My local branch is better that gerrit's one. lol | |
| 18:40:15 | andreykurilin | bauzas: yes, 2.41 is latest supported api version | |
| 18:40:37 | bauzas | andreykurilin: okay, so it's confirmed, I can't use your SHA1 | |
| 18:40:44 | bauzas | andreykurilin: http://git.openstack.org/cgit/openstack/python-novaclient/log/ | |
| 18:41:02 | bauzas | andreykurilin: compared to https://review.openstack.org/#/c/425822/1//COMMIT_MSG@1 | |
| 18:41:15 | bauzas | andreykurilin: your parent patch is not HEAD | |
| 18:41:44 | andreykurilin | bauzas: It can be fixed easily by pushing "Rebase" button via gerrit | |
| 18:42:00 | bauzas | don't | |
| 18:42:13 | bauzas | leave the patch be merged and I'll pick the merge commit sha1 | |
| 18:42:19 | andreykurilin | np :) | |
| 18:44:41 | bauzas | andreykurilin: FYI https://review.openstack.org/#/c/425827/ | |
| 18:44:57 | bauzas | andreykurilin: it's pointing on the current HEAD | |
| 18:45:38 | bauzas | andreykurilin: in case I'm not there, just ping matt with that one and he'll update inline the change | |
| 18:45:42 | snikitin | alex_xu, mriedem: If you are around could you too look at diagnostics patches, please? Today at one moment I had five +2 on my three patches, but had lost all of them because of rebases :( https://review.openstack.org/#/c/394480/ | |
| 18:45:55 | bauzas | or if we are too late, just remove the mention of DNM and leave that be cut | |
| 18:46:27 | andreykurilin | bauzas: got it, thanks | |
| 18:46:35 | bauzas | back in 1 hour-ish | |
| 18:48:43 | openstackgerrit | OpenStack Proposal Bot proposed openstack/nova: Updated from global requirements https://review.openstack.org/424828 | |
| 18:53:08 | dansmith | bauzas: several things failing on both patches | |
| 18:56:24 | openstackgerrit | Sergey Nikitin proposed openstack/nova: Change parent of diagnostics objects to NovaObject class https://review.openstack.org/394480 | |
| 18:56:25 | openstackgerrit | Sergey Nikitin proposed openstack/nova: Standardization of VM diagnostics info API. https://review.openstack.org/355540 | |
| 18:56:27 | openstackgerrit | Sergey Nikitin proposed openstack/nova: Completed implementation of instance diagnostics for Xen https://review.openstack.org/399613 | |
| 19:07:08 | openstackgerrit | Sergey Nikitin proposed openstack/nova: Change parent of diagnostics objects to NovaObject class https://review.openstack.org/394480 | |
| 19:07:09 | openstackgerrit | Sergey Nikitin proposed openstack/nova: Standardization of VM diagnostics info API. https://review.openstack.org/355540 | |
| 19:07:11 | openstackgerrit | Sergey Nikitin proposed openstack/nova: Completed implementation of instance diagnostics for Xen https://review.openstack.org/399613 | |
| 19:08:06 | cburgess | Is there an API call to get the contents of the block_device_mapping table for server or volume ID? | |
| 19:08:11 | cburgess | Or is that purely internal? | |
| 19:10:22 | artom | cburgess, the deprecated /os-volumes endpoint is probably the closest thing to what you want | |
| 19:10:22 | artom | http://developer.openstack.org/api-ref/compute/?expanded=list-volumes-detail | |
| 19:11:31 | cburgess | Tried that.. tragically it doesn't seem to work. | |
| 19:11:36 | cburgess | I'll take another look at it. | |
| 19:12:05 | artom | cburgess, define not working... | |
| 19:12:16 | artom | Since it's deprecated, you need to cap your client's compute api version to 2.36 | |
| 19:12:49 | cburgess | This AZ happens to be liberty so its capped at 2.12 | |
| 19:14:03 | cburgess | As for not working... let me rerun the command. | |
| 19:16:30 | cburgess | Ahhh ok it works. | |
| 19:16:59 | cburgess | But it does't expose the connection_info field from the block_device_mapping table. | |
| 19:17:16 | cburgess | At least not that I can see | |
| 19:17:31 | cburgess | I was previously passing a volume rather then an attachment id that was what was not working. | |
| 19:17:39 | dansmith | bauzas: os_region_name is apparently unset on the grenade job so your computes are all not starting | |
| 19:18:51 | dansmith | which makes sense I guess without the other grenade change because we're running with the default config from newton | |
| 19:18:58 | dansmith | so your middle patch needs the dep from the last one | |
| 19:25:11 | openstackgerrit | Dan Smith proposed openstack/nova: Scheduler calling the Placement API https://review.openstack.org/417961 | |
| 19:25:12 | openstackgerrit | Dan Smith proposed openstack/nova: Block starting compute unless placement conf is provided https://review.openstack.org/425806 | |
| 19:25:23 | dansmith | bauzas: made the caching change, changed the dep to the middle patch, and addressed cdent's comments ^ | |
| 19:25:30 | artom | cburgess, well, if you have a use case for it, you can always propose a spec for a new API microversion that does what you want :) | |
| 19:25:38 | cdent | thanks dansmith | |
| 19:26:56 | cburgess | artom I might later. RIght now just trying to fix some corruption that got into the DB (probably from some long ago fixed bug). Might make sense in the future but for now I'll just load up MySQLdb to find what I need. | |
| 19:28:33 | artom | cburgess, aha! Well, good luck in your quest ;) | |
| 19:29:16 | artom | Actually for that use case, straight up DB access is easier, no? You can pipe it to grep and sed and awk and stuff to find what you're looking for easier | |
| 19:31:09 | cburgess | artom Well I'm writting a python script to do what I need but yeah I suppose it is. I just prefer to avoid having to muck about in the DB if I can avoid it. | |
| 19:38:36 | andymccr | evenin' nova - on the placement api - is there a way to specify which interface (internal/public/admin) will be used to connect to the placement api? i can't see a conf option for that, and it seems to default to public. | |
| 19:39:58 | dansmith | andymccr: it defaults to however you route to that host | |
| 19:40:05 | dansmith | I mean, | |
| 19:40:19 | dansmith | we don't bind to an interface, we just connect to that host, so normal routing applies :) | |
| 19:41:41 | andymccr | dansmith: i mean the "keystone endpoint interface" thats picked - so it looks like its using the publicURL I'm wondering if I can get it to use the internalURL- | |
| 19:43:23 | dansmith | andymccr: I don't know enough about the keystoneclient bits, but .. if you don't set a publicurl maybe? | |