| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2021-02-15 | |||
| 19:30:07 | sean-k-mooney | using raw/qcow does not percluded you mounting the nova state dir on an lvm volumn seperatly | |
| 19:33:15 | kukacz | sean-k-mooney: you are right, it does not. I have just defined myself different test scenarios originally, not including this one | |
| 19:34:47 | sean-k-mooney | my home openstack has my nova directory in an lvm vg on a raid 6 disk array with an optane ssd as a cache via bcache | |
| 19:35:08 | kukacz | with the current knowledge, it seems the best option. supposing I can get to slightly comparable performance | |
| 19:36:29 | sean-k-mooney | in your case you dont need bcach since its all ssd but for me its nice to hide some of the latency of HDDs | |
| 19:36:50 | sean-k-mooney | anyway o/ | |
| 19:40:11 | kukacz | of course. I'm also trying to keep the configuration as simple as possible. avoiding any extra layers. that's why I'm so enthusiastic of the idea of direct LVM usage | |
| 19:42:12 | kukacz | if I put filesystem (/var/lib/nova/instances) on top of LVM, it will not partially-tolerate disk failure anymore. that could only be achieved with direct Nova-LVM integration | |
| 19:43:35 | kukacz | seems that I'll need to accept a compromise | |
| #openstack-nova - 2021-02-16 | |||
| 09:03:08 | lyarwood | Morning | |
| 09:03:42 | xarlos | o/ | |
| 09:13:40 | lyarwood | gibi / stephenfin ; thanks again for the reviews yesterday, just going over the initial changes now, if you have time this morning I'd appreciate any reviews on the latter changes. | |
| 09:25:05 | stephenfin | ack | |
| 09:26:15 | openstackgerrit | Belmiro Moreira proposed openstack/nova master: Change API unexpected exception message https://review.opendev.org/c/openstack/nova/+/628061 | |
| 09:37:04 | belmoreira | Hello, just had another user confused with the API error: 'Unexpected API Error. Please report this at http://bugs.launchpad.net/nova/ and attach the Nova API log if possible.' and I remember that we already discussed it. I would like to revive it. Do you still think that a config option is the way to go? Because I can't see how this specific error message is good for any cloud provider. | |
| 09:37:08 | belmoreira | https://review.opendev.org/c/openstack/nova/+/628061 | |
| 09:37:37 | belmoreira | let me know your comments. I would like to work on it. | |
| 09:41:39 | lyarwood | belmoreira: +1 to making it configurable somehow | |
| 09:42:02 | lyarwood | belmoreira: we already have vendor specific configurables ./me checks | |
| 09:45:26 | lyarwood | belmoreira: https://github.com/openstack/nova/blob/master/nova/version.py is what I was thinking of | |
| 09:45:40 | lyarwood | belmoreira: adding some form of bug reporting URL there that defaults to http://bugs.launchpad.net/nova/ would be nice | |
| 09:46:26 | lyarwood | belmoreira: and you could overwrite that in your deployment to point at your support portal etc | |
| 09:47:46 | lyarwood | https://github.com/rdo-packages/nova-distgit/blob/73a5163768f8d093867e1f7d9ea1129e4a0a839a/openstack-nova.spec#L525-L531 is how we do this downstream in RDO and OSP | |
| 09:48:07 | lyarwood | well, overwrite the vendor, product and package | |
| 09:48:24 | lyarwood | and tbh we could also use a bug_tracker configurable that points to our support portal there | |
| 09:48:37 | belmoreira | lyarwood thanks, let me check | |
| 09:54:12 | lyarwood | gah these gerrit comment threads are confusing | |
| 09:58:29 | lyarwood | stephenfin: https://review.opendev.org/c/openstack/nova/+/769548/10/nova/tests/unit/virt/test_machine_type_utils.py - stupid question, do you mean drop the empty line? | |
| 09:58:55 | stephenfin | oh, yeah | |
| 09:59:04 | stephenfin | we haven't switched to unittest.mock yet | |
| 09:59:15 | lyarwood | right I was about to say | |
| 09:59:22 | lyarwood | I thought I had missed something with unittest.mock | |
| 09:59:29 | lyarwood | cool | |
| 10:54:29 | gibi | lyarwood: ack, I'm slowly reviewing the rest of the machine_type series but there is a lot of distraction around me at the moment | |
| 10:55:10 | lyarwood | gibi: np at all, take your time I appreciate it all the same :) | |
| 10:55:25 | lyarwood | >< | |
| 10:55:34 | lyarwood | I'd be made if it wasn't so cute | |
| 10:55:36 | lyarwood | mad* | |
| 10:55:52 | stephenfin | bauzas: Want to take a quick look at https://review.opendev.org/c/openstack/os-traits/+/759878 so we can cut an os-traits release? | |
| 10:56:40 | bauzas | stephenfin: currently fighting with the odds of _IntegratedTestBase but sure can look :) | |
| 10:57:07 | gibi | lyarwood: you won | |
| 10:57:20 | stephenfin | gibi: I've proposed an os-traits 2.5.0 release https://review.opendev.org/c/openstack/releases/+/775804 and think we could merge it without https://review.opendev.org/c/openstack/os-traits/+/759878 | |
| 10:57:24 | stephenfin | seeing as releases are cheap | |
| 10:57:25 | bauzas | fwiw we're 99% on par with ProviderUsageBaseTestCase but I hit the 1% left :) | |
| 10:58:03 | bauzas | yeah we can release it | |
| 10:58:38 | bauzas | and here I'm ok with playing the same team card | |
| 10:58:44 | bauzas | it's just a trait | |
| 10:59:09 | stephenfin | thanks | |
| 10:59:17 | gibi | hm so https://review.opendev.org/c/openstack/os-traits/+/759878 is now approved | |
| 10:59:23 | gibi | then I guess there will be a new release patch | |
| 10:59:41 | stephenfin | Yeah, might as well | |
| 11:00:01 | gibi | (I would have approved that os-trait patch by myself but bauzas was faster) | |
| 11:00:14 | gibi | stephenfin: I think the os-trait merge should not take long time | |
| 11:00:18 | bauzas | gibi: I added a comment about it | |
| 11:00:25 | bauzas | about the same company one | |
| 11:00:50 | bauzas | gibi: agreed about releasing it quickly | |
| 11:01:06 | gibi | bauzas: no worries about the same company rule | |
| 11:02:36 | gibi | stephenfin: is it OK for you to wait for this patch to merge and then release the os-traits? | |
| 11:02:45 | stephenfin | yup, wfm | |
| 11:02:46 | gibi | I think we can do that today | |
| 11:07:13 | lyarwood | so I've not a single gerrit review email this morning, odd. | |
| 11:09:13 | lyarwood | gah nvm it's offlineimap *again* | |
| 11:09:23 | lyarwood | I should really switch to gmail and give up with mutt etc | |
| 11:11:43 | openstackgerrit | Merged openstack/os-traits master: Add COMPUTE_EPHEMERAL_ENCRYPTION tratis https://review.opendev.org/c/openstack/os-traits/+/759878 | |
| 11:14:31 | gibi | that was faaaast ^^ | |
| 11:14:55 | stephenfin | yup, https://review.opendev.org/c/openstack/releases/+/775804 updated | |
| 11:15:01 | gibi | looking | |
| 11:15:53 | gibi | done | |
| 11:17:30 | stephenfin | thanks | |
| 11:52:03 | kashyap | lyarwood: OfflineIMAP isn't being ported to Python-3 :-( It's Python-2-only | |
| 11:52:16 | kashyap | lyarwood: I have been planning to migrate to `mbsync`, and keep delaying it | |
| 11:52:34 | stephenfin | lyarwood: Are you okay with me merging https://review.opendev.org/c/openstack/nova/+/775415 into https://review.opendev.org/c/openstack/nova/+/769920/5 ? | |
| 11:52:40 | kashyap | [I know, Python version isn't the problem in your case now] | |
| 11:52:41 | stephenfin | I need that work for the Secure Boot stuff also | |
| 11:52:52 | stephenfin | I can keep them separate too if you'd prefer | |
| 11:53:18 | lyarwood | stephenfin: yeah merge/nuke it given you've fixed things up in your change | |
| 11:53:40 | lyarwood | stephenfin: to be clear, close mine out. | |
| 11:55:31 | kashyap | stephenfin: Thanks for the review here; /me goes to read: https://review.opendev.org/c/openstack/nova/+/774240 | |
| 11:56:00 | kashyap | stephenfin: lyarwood: My capacity this week is significantly reduced due to some personal event :-( and next week I'm on PTO | |
| 11:56:13 | kashyap | So I'll be somewhat less in action; but I'd like to address the feedback on the CPU stuff | |
| 11:56:57 | lyarwood | ack np, I can help push things along while you're out this week, I'm also out Thursday/Friday next week FWIW. | |
| 11:57:54 | kashyap | Thank you; the CPU stuff is in shape, modulo Stephen's comments. (And just trying to close out some more pressing downstream stuff, too) | |
| 11:58:20 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Add 'nova.virt.libvirt.utils.get_arch' stub to fixture https://review.opendev.org/c/openstack/nova/+/774904 | |
| 11:58:20 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Remove unnecessary mock https://review.opendev.org/c/openstack/nova/+/774905 | |
| 11:58:21 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Poison os.uname https://review.opendev.org/c/openstack/nova/+/775415 | |
| 11:58:26 | kashyap | I'm still here this week, but functioning less than 100%. (Next week I'm off) | |
| 11:58:26 | stephenfin | b****** | |
| 11:58:43 | stephenfin | that shouldn't have rebased the whole series :-( | |
| 11:59:27 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Add 'nova.virt.libvirt.utils.get_arch' stub to fixture https://review.opendev.org/c/openstack/nova/+/774904 | |
| 11:59:28 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Remove unnecessary mock https://review.opendev.org/c/openstack/nova/+/774905 | |
| 11:59:28 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Poison os.uname https://review.opendev.org/c/openstack/nova/+/775415 | |
| 12:00:29 | lyarwood | -R ftw | |
| 12:01:12 | stephenfin | nah, I've a bad habit of using 'git rebase -i master' to do any history rewriting work | |
| 12:01:21 | stephenfin | vs. git rebase -i HEAD~2 or whatever I needed there | |
| 12:05:47 | lyarwood | huh does that edit the entire series from origin/master/HEAD? | |
| 12:06:24 | stephenfin | rebases and lets me do what I want to the entire series | |
| 12:07:05 | lyarwood | ah it rebases onto master, so if that changes the entire series does | |
| 12:07:14 | stephenfin | yup | |
| 12:07:20 | lyarwood | yeah I just use -i HEAD~$n | |
| 12:07:37 | stephenfin | a wiser approach | |