| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-10-29 | |||
| 19:19:55 | mnaser | Which I assume means libvirt will be taking care of not making us do bad things | |
| 19:32:20 | mriedem | mnaser: comments inline | |
| 19:32:21 | mriedem | mostly just nits | |
| 19:36:22 | mriedem | cfriesen: you're talking about this right? https://github.com/openstack/nova/blob/a0eacbf7fff60282007ddca705ef7331e8a4a6f8/nova/virt/libvirt/migration.py#L83 | |
| 19:36:41 | mriedem | yeah that's what i was referring to earlier - before _update_vif_xml we'd just always send the source vif xml unchanged | |
| 19:37:25 | cfriesen | mriedem: yes | |
| 19:44:35 | mnaser | mriedem: cfriesen thanks for the reviews both, i tried to address but in case it wasn't clear in terms of comments, please let me know | |
| 19:44:44 | openstackgerrit | Mohammed Naser proposed openstack/nova master: Add tests for bug #1800511 https://review.openstack.org/614004 | |
| 19:44:44 | openstack | bug 1800511 in OpenStack Compute (nova) "VMs with vif_type bridge/tap started before Rocky upgrade cannot be live migrated" [High,In progress] https://launchpad.net/bugs/1800511 - Assigned to Mohammed Naser (mnaser) | |
| 19:44:44 | openstackgerrit | Mohammed Naser proposed openstack/nova master: libvirt: Avoid setting MTU during live migration if unset https://review.openstack.org/614008 | |
| 19:50:50 | cfriesen | mnaser: much better, one little comment | |
| 19:52:07 | mnaser | cfriesen: i guess i wanted to make sure that a normal reboot wouldn't do the trick (because the qemu-kvm process doesnt get restarted) | |
| 19:52:13 | mnaser | but i guess the follow up comment clears it up | |
| 19:52:47 | openstackgerrit | Mohammed Naser proposed openstack/nova master: libvirt: Avoid setting MTU during live migration if unset https://review.openstack.org/614008 | |
| 19:52:50 | mnaser | cfriesen: voila | |
| 19:52:51 | cfriesen | mnaser: a "nova reboot" without the "hard" option will reset things | |
| 19:52:59 | mnaser | really | |
| 19:53:04 | mnaser | i thought i tried to do an acpi reboot | |
| 19:54:20 | cfriesen | we try an ACPI shutdown | |
| 19:54:31 | cfriesen | then if that doesn't work we do a hard shutdown | |
| 19:54:39 | mnaser | aaaah | |
| 19:54:40 | mnaser | gotcha | |
| 19:54:44 | mnaser | so the reboot always restarts the process | |
| 19:54:54 | mnaser | cool, TIL! | |
| 19:56:11 | mriedem | pep8 is going to fail on https://review.openstack.org/#/c/614004/ | |
| 19:56:57 | mnaser | ah shit | |
| 19:57:42 | mnaser | ill run that locally | |
| 19:58:03 | mriedem | +2 on the top one | |
| 20:21:12 | openstackgerrit | Mohammed Naser proposed openstack/nova master: Add tests for bug #1800511 https://review.openstack.org/614004 | |
| 20:21:12 | openstackgerrit | Mohammed Naser proposed openstack/nova master: libvirt: Avoid setting MTU during live migration if unset https://review.openstack.org/614008 | |
| 20:21:13 | openstack | bug 1800511 in OpenStack Compute (nova) "VMs with vif_type bridge/tap started before Rocky upgrade cannot be live migrated" [High,In progress] https://launchpad.net/bugs/1800511 - Assigned to Mohammed Naser (mnaser) | |
| 20:21:30 | mnaser | mriedem: fixed your nit too if you feel like readding your +2 and cfriesen +1 :) | |
| 20:24:28 | mriedem | +Ws | |
| 20:24:30 | mriedem | oops | |
| 20:24:32 | mriedem | +2s | |
| 20:24:46 | mriedem | time to play the "find another core" game | |
| 20:36:22 | mnaser | mriedem: I guess all the IBM ones are busy today :p | |
| 20:38:34 | mriedem | too soon | |
| 20:42:25 | mnaser | It’s been quiet here today | |
| 20:42:41 | mnaser | And it’s getting late so I might have to settle for an EU core tomorrow :p | |
| 20:45:47 | melwitt | I can take a look. things have been a bit chaotic today | |
| 20:48:28 | mriedem | the one day sean-k-mooney isn't working 20 hours | |
| 20:49:16 | mriedem | mnaser: you've deployed this patch now and verified it fixes your upgrade issue? | |
| 20:49:30 | melwitt | do we want sean-k-mooney to test this first? | |
| 20:49:42 | mriedem | i think mnaser is testing it in prod | |
| 20:49:56 | mnaser | I have not tested in the environment to be honest | |
| 20:50:08 | mnaser | I can do that now tho | |
| 20:50:15 | mnaser | Gimme a few | |
| 20:50:28 | dansmith | seems pretty clear from the bug that this is the right thing to do | |
| 20:50:38 | melwitt | ok, asked in case that's why you mentioned sean-k-mooney | |
| 20:50:50 | dansmith | obviously a manual test would be nice | |
| 20:51:06 | dansmith | I dropped a +W on the top meaty one.. I can put an upgradeable +1 on the bottom one | |
| 21:03:25 | mnaser | (taking a bit of time, trying to find VMs that have been launched before which are testing) | |
| 21:14:33 | mnaser | dansmith: mriedem melwitt i've tested this one now locally on a vm that confirmed had no mtu set (virsh dumpxml, /proc/pid/cmdline had no host_mtu), live migrated with patch and migration completed succesfully | |
| 21:18:52 | mriedem | i mentioned sean-k-mooney because he added the mtu stuff for bridge/tap types | |
| 21:19:04 | mriedem | and is familiar with this live migration _update_vif_xml stuff | |
| 21:20:48 | melwitt | ack | |
| 21:23:51 | mriedem | while we're all here and having fun, we need to do a rocky release for several upgrade related issues, but need https://review.openstack.org/#/q/topic:bug/1798188+(status:open+OR+status:merged) as part of that, | |
| 21:23:58 | mriedem | and will also want the fix that mnaser just pushed | |
| 21:24:02 | mriedem | mnaser: are you working backports? | |
| 21:24:18 | efried | mriedem: Allocation ratio fix will need to be part of --^ ? | |
| 21:24:32 | mriedem | efried: i don't think that is critical | |
| 21:24:36 | mriedem | annoying, yes | |
| 21:24:39 | mriedem | but not critical | |
| 21:24:39 | efried | okay | |
| 21:28:12 | melwitt | mriedem: I was debating how to update the reno as mentioned by mnaser on the review | |
| 21:29:18 | melwitt | because once all compute hosts are rocky, should be OK to stop running nova-consoleauth with the workaround BUT you also have to make sure you deploy your console proxies per cell and you will still experience all consoles being reset that the point you stop using the workaround (if not all auths already expired) | |
| 21:29:58 | melwitt | it's kind of complicated so I was feeling stuck on whether the reno is conveying all of that well enough | |
| 21:31:21 | mriedem | well, it says use the workaround while you're doing rolling upgrade, so not all computes are rocky yet, as mohammed pointed out, | |
| 21:31:51 | mriedem | and then there is the "Operators may unset the configuration option when:" section which says once you're upgraded (all hosts are Rocky), and stuff is expired, you can drop the workaround, | |
| 21:32:01 | mriedem | if you need to add "and you're running console proxy per cell" to that list, then just add that | |
| 21:32:38 | melwitt | yeah, ok. thanks. I've felt so mired in it that I was having a hard time thinking about it | |
| 21:32:49 | mriedem | i left comments, | |
| 21:33:17 | mriedem | but i think it's just (1) describing what rolling upgrade means and (2) add the part about console proxy being deployer per cell before unsetting the workaround | |
| 21:34:02 | mriedem | we have checks for the former in the nova-status upgrade check, but not the latter, but i'm not sure if it's possible to check the latter via nova-status is it? | |
| 21:39:46 | openstackgerrit | Matt Riedemann proposed openstack/nova master: WIP: Cross-cell resize https://review.openstack.org/603930 | |
| 21:39:46 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Add Destination.cross_cell_move field https://review.openstack.org/614035 | |
| 21:39:47 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Add InstanceAction/Event create() method https://review.openstack.org/614036 | |
| 21:39:47 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Change HostManager to allow scheduling to other cells https://review.openstack.org/614037 | |
| 21:39:57 | melwitt | mriedem: I don't think so. the only possible way to check is if the nova.conf on the console proxy host has [database]connection configured (implying access to a cell database) | |
| 21:42:08 | mnaser | mriedem: sorry checking things in and out, let me push up backport | |
| 21:42:28 | mnaser | I wonder if the UI adds the cherry picked from if it hasn’t merged yet | |
| 21:42:32 | mriedem | it does not | |
| 21:42:46 | mnaser | Boo. Alright | |
| 21:42:52 | mnaser | This needs to go back to rocky or further? | |
| 21:43:05 | mriedem_away | rocky | |
| 21:43:09 | mnaser | ok cool | |
| 21:43:36 | mriedem_away | _update_vif_xml was added in rocky, | |
| 21:44:10 | mriedem_away | setting the mtu was backported to queens. if that could fail for different reasons, like live migrating a guest from source host which set mtu to something and then to a dest host that is not new enough to support that mtu config, well that's a different issue | |
| 21:44:45 | mnaser | mriedem_away: most likely not, this feature was added a while back ago | |
| 21:44:51 | mnaser | let me verify that | |
| 21:45:01 | mriedem_away | libvirt 3.3 | |
| 21:45:27 | mnaser | oh yes only that got support for it | |
| 21:45:39 | mriedem_away | but our minimum libvirt support is 1.3.1 | |
| 21:45:56 | mnaser | so unless someone is running 1.3.1 alongside 3.3.0 | |
| 21:46:08 | mriedem_away | i think this might catch it anyway https://github.com/openstack/nova/blob/master/nova/conductor/tasks/live_migrate.py#L235 | |
| 21:46:37 | mriedem_away | alright really leaving now | |
| 21:46:42 | openstackgerrit | Mohammed Naser proposed openstack/nova stable/rocky: Add tests for bug #1800511 https://review.openstack.org/614040 | |
| 21:46:42 | openstack | bug 1800511 in OpenStack Compute (nova) "VMs with vif_type bridge/tap started before Rocky upgrade cannot be live migrated" [High,In progress] https://launchpad.net/bugs/1800511 - Assigned to Mohammed Naser (mnaser) | |
| 21:46:42 | openstackgerrit | Mohammed Naser proposed openstack/nova stable/rocky: libvirt: Avoid setting MTU during live migration if unset https://review.openstack.org/614041 | |