| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-02-20 | |||
| 15:30:37 | mriedem | mnaser: i can update it quick | |
| 15:30:54 | mriedem | we'll lose mel's +2 | |
| 15:31:39 | jaypipes | mriedem, dansmith, bauzas, melwitt: I'm good with efried's https://review.openstack.org/#/c/546009/. seems like a very simple API enhancement that will yield positive performance impacts. | |
| 15:31:40 | mnaser | mriedem: always bothered me that updating commit messages drops +2's but ehh.. if you think its fine and can convince someone to review and +A then even better :p | |
| 15:32:04 | mnaser | i just want to follow the patch through gates and get it backported (hopefully cleanly) | |
| 15:33:29 | openstackgerrit | Chris Dent proposed openstack/nova master: WIP: experiment with setting placement error codes https://review.openstack.org/546177 | |
| 15:33:53 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Clean up ports and volumes when deleting ERROR instance https://review.openstack.org/340614 | |
| 15:33:54 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Detach volumes when deleting a BFV server pre-scheduling https://review.openstack.org/545132 | |
| 15:33:54 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Add functional recreate test of deleting a BFV server pre-scheduling https://review.openstack.org/545123 | |
| 15:34:04 | mriedem | ah shit | |
| 15:34:11 | mriedem | bauzas: ^ can you re-approve those bottom 2? | |
| 15:34:31 | efried | mriedem: Dangit, where's that placement doc with the "creating a new microversion" checklist? | |
| 15:34:53 | mriedem | https://docs.openstack.org/nova/latest/search.html | |
| 15:35:10 | bauzas | mriedem: sure thing | |
| 15:35:34 | cdent | efried: this one? https://docs.openstack.org/nova/latest/contributor/placement.html | |
| 15:35:58 | efried | cdent, mriedem: yup, thank you. | |
| 15:36:00 | openstackgerrit | Surya Seetharaman proposed openstack/nova master: Unmap compute nodes when deleting host mappings in delete cell operation https://review.openstack.org/542964 | |
| 15:36:35 | openstack | Launchpad bug 1740025 in OpenStack Compute (nova) "LVM Sparse Volumes broken" [Undecided,New] | |
| 15:36:35 | bauzas | mriedem: considering https://bugs.launchpad.net/nova/+bug/1740025 | |
| 15:36:58 | bauzas | WDYT about deprecating the LVM sparse stuff ? | |
| 15:37:06 | efried | cdent: I wouldn't be offended if that linked off of either/both of https://developer.openstack.org/api-ref/placement/ https://docs.openstack.org/nova/latest/user/placement.html | |
| 15:37:26 | mriedem | bauzas: don't know anything about it, but i think cfriesen does | |
| 15:37:26 | efried | certainly the latter | |
| 15:38:01 | mriedem | efried: user docs don't really need to link to dev docs | |
| 15:38:23 | efried | Guess it depends on your definition of "user". | |
| 15:38:36 | mriedem | user != contributor writing code etc | |
| 15:38:38 | mriedem | imo | |
| 15:38:48 | efried | I would think there's enough crossover that such a link wouldn't hurt. | |
| 15:39:20 | efried | but whatevs. Now it's permanently open in a tab for me, so I don't care anymore. | |
| 15:39:24 | cdent | A "if you're curious about this stuff, seem dev notes" kind of link on the user might be useful for some (but could also confuse some) but having it on the api-ref: meh | |
| 15:39:51 | dansmith | mriedem: can you drop your -1 on this at least? https://review.openstack.org/#/c/543580/ | |
| 15:39:58 | mriedem | or just https://docs.openstack.org/nova/latest/contributor/ | |
| 15:40:12 | mriedem | dansmith: that's the least i can do | |
| 15:40:20 | mriedem | dansmith: so those grenade results were just due to old setup? | |
| 15:40:42 | bauzas | mriedem: okay, fair enough | |
| 15:40:43 | dansmith | mriedem: yeah, from doing a pike->master upgrade | |
| 15:41:02 | bauzas | mriedem: the problem I see with LVM sparse LVs is that I guess nothing was checked | |
| 15:41:14 | dansmith | mriedem: the auto selection was choosing the right version, and master was rightly saying "I don't speak that version, yo" | |
| 15:41:37 | mriedem | hmm | |
| 15:41:42 | mriedem | with that 90s rap slang? | |
| 15:41:55 | dansmith | mriedem: obviously | |
| 15:42:07 | mriedem | ok; it's in a tab | |
| 15:42:08 | dansmith | mriedem: "this is how we dooooo, RPC versioning" | |
| 15:42:35 | mriedem | oh btw you might want to punch any red hat infra people about https://review.openstack.org/#/c/546183/ | |
| 15:42:44 | mriedem | because it turns out we have http://grafana.openstack.org/dashboard/db/ceph-failure-rate | |
| 15:42:47 | mriedem | but it's been broken since october | |
| 15:43:18 | dansmith | mriedem: I was taught not to punch people | |
| 15:43:26 | mriedem | wet willy some red hat infra people | |
| 15:43:29 | dansmith | mriedem: unless they were wrong or said something I didn't like | |
| 15:43:31 | dansmith | hah, eww | |
| 15:47:42 | hrw | https://marcin.juszkiewicz.com.pl/2018/02/19/hotplug-in-vm-easy-to-say/ - a story about https://review.openstack.org/#/c/545034/ patch (which waits for reviewers) | |
| 15:50:32 | openstack | Launchpad bug 1740025 in OpenStack Compute (nova) "LVM Sparse Volumes broken" [Undecided,Won't fix] | |
| 15:50:32 | bauzas | mriedem: okay, so I did set Wontfix for https://bugs.launchpad.net/nova/+bug/1740025 given my last comment | |
| 15:54:03 | mriedem | bauzas: if the fix is to deprecate the sparse_logical_volumes option, then it shouldn't really be 'wont fix' | |
| 15:54:15 | mriedem | unless you're just gonna related-bug that thing | |
| 15:55:01 | bauzas | mriedem: I'll deprecate the conf opt by passing a related-bug, yes | |
| 15:55:27 | bauzas | Wontfix is exactly for like "sorry, we know it doesn't work, but we don't plan to fix it" | |
| 15:55:51 | bauzas | and that bug is the exact reason why we should deprecate that option | |
| 15:56:06 | mriedem | bauzas: did you take a look at https://review.openstack.org/#/c/545132/ ? | |
| 15:56:16 | mriedem | you +W'ed the regression recreate test | |
| 15:56:42 | mriedem | mnaser: you should be able to start backports for that series | |
| 16:01:32 | bauzas | mriedem: not yet, because I need to think about it | |
| 16:02:14 | bauzas | ie. if the call is made in the right place | |
| 16:02:20 | bauzas | and if we're not missing anything | |
| 16:02:29 | bauzas | the change itself is easy to review | |
| 16:04:06 | mnaser | mriedem: i'll start doing that | |
| 16:05:16 | openstack | Launchpad bug 1739735 in OpenStack Compute (nova) "boot baremetal server which has multi-interface will randomly choose baremetal server interface" [Undecided,New] | |
| 16:05:16 | bauzas | mgoddard: saw your comment on https://bugs.launchpad.net/nova/+bug/1739735 looks like it's a new feature, right? | |
| 16:08:25 | cfriesen | mriedem: bauzas: I haven't used the LVM sparse stuff, we do use thinly-provisioned LVM internally. | |
| 16:08:48 | bauzas | heh | |
| 16:09:06 | bauzas | I'm pretty sure that if you used it, you'd discovered problems: p | |
| 16:09:36 | bauzas | it's a canary in a coal mine :p | |
| 16:10:17 | bauzas | artom: around ? | |
| 16:10:37 | bauzas | artom: do you think that https://bugs.launchpad.net/nova/+bug/1739735 is solvable by http://specs.openstack.org/openstack/nova-specs/specs/pike/implemented/virt-device-tagged-attach-detach.html ? | |
| 16:10:38 | openstack | Launchpad bug 1739735 in OpenStack Compute (nova) "boot baremetal server which has multi-interface will randomly choose baremetal server interface" [Undecided,New] | |
| 16:10:45 | bauzas | artom: ie. using the ironic driver | |
| 16:10:47 | mriedem | dansmith: in https://review.openstack.org/#/c/543580/ - why do you remove defaults for kwargs in some methods but not others? | |
| 16:11:43 | dansmith | mriedem: because those are things that were added as optional just for compatibility when adding a new param that really should be passed all the time | |
| 16:11:44 | cfriesen | bauzas: we have internal patches to add support for thin lvm to nova. also to enable cold-migration and resize for LVM-backed instances | |
| 16:12:20 | dansmith | mriedem: if it was a thing we added in 4.x and expect to always pass, but made it optional so the tests would pass without having to touch it, I made it required here and fixed up the tests | |
| 16:12:31 | dansmith | mriedem: or at least, that was the intent | |
| 16:12:36 | bauzas | cfriesen: see https://bugs.launchpad.net/nova/+bug/1740025/comments/3 | |
| 16:12:37 | openstack | Launchpad bug 1740025 in OpenStack Compute (nova) "LVM Sparse Volumes broken" [Undecided,Won't fix] | |
| 16:12:49 | bauzas | cfriesen: I seriously consider deprecating that possibility | |
| 16:12:56 | bauzas | at least the config option | |
| 16:13:04 | mriedem | dansmith: hmm, well not all are like that, like rebuild_instance shouldn't require host/node | |
| 16:13:23 | dansmith | mriedem: which line? | |
| 16:13:37 | mriedem | https://review.openstack.org/#/c/543580/1/nova/compute/rpcapi.py@721 | |
| 16:14:20 | bauzas | jroll: trying to get Ironic's eyes for https://bugs.launchpad.net/nova/+bug/1739735 | |
| 16:14:21 | openstack | Launchpad bug 1739735 in OpenStack Compute (nova) "boot baremetal server which has multi-interface will randomly choose baremetal server interface" [Undecided,New] | |
| 16:14:23 | cfriesen | bauzas: I'd be fine with that. :) One main reason we added support for thin volumes was that zeroing out the volumes was incredibly disk-intensive. By using thin we only zero out the blocks as they are allocated. | |
| 16:14:25 | mriedem | dansmith: nvm, | |
| 16:14:35 | mriedem | i was thinking of the conductor / compute task rpcapi | |
| 16:14:41 | bauzas | cfriesen: cool | |
| 16:14:43 | jroll | bauzas: will look / pass along | |
| 16:14:43 | mriedem | conductor will always pass host/node to the compute rebuild_instance | |
| 16:14:55 | bauzas | jroll: mgoddard provided a comment | |
| 16:15:08 | mriedem | even if it's just None | |
| 16:15:11 | bauzas | jroll: but I'm unclear if/how that makes that bug fixable | |
| 16:15:17 | dansmith | mriedem: it's just super hard to reason about a lot of these calls where everything is optional, except not really | |