| opendevreview | Merged openstack/nova master: Run functional test with threading mode https://review.opendev.org/c/openstack/nova/+/979850 | 02:26 |
|---|---|---|
| *** benj_8 is now known as benj_ | 06:15 | |
| opendevreview | ChungWon Lee proposed openstack/nova master: Add regression test for bug #2134375 https://review.opendev.org/c/openstack/nova/+/984900 | 08:16 |
| tafkamax | Hi, we have a question regarding live-migrate. Why do some nodes live-migrate really fast and some take many minutes up to 20 or so. | 08:34 |
| tafkamax | I know it copies over RAM. But an VM with 8gb took many minutes and a vm with 24gb took 40 seconds or so. | 08:35 |
| tafkamax | It feels very random. | 08:35 |
| tafkamax | Where can I look for how this works and I can find a deterministic way to see why it varies so much? | 08:39 |
| *** jgilaber_ is now known as jgilaber | 08:54 | |
| opendevreview | Kamil Sambor proposed openstack/nova master: Deprecate eventlet concurrency mode https://review.opendev.org/c/openstack/nova/+/987331 | 09:55 |
| opendevreview | Elod Illes proposed openstack/nova stable/2025.1: fix: device_by_alias should respect config type https://review.opendev.org/c/openstack/nova/+/984921 | 10:58 |
| *** mhen_ is now known as mhen | 12:05 | |
| *** erlon3 is now known as erlon | 12:45 | |
| opendevreview | Stephen Finucane proposed openstack/os-vif master: Drop support for Python 3.10 https://review.opendev.org/c/openstack/os-vif/+/987500 | 13:17 |
| opendevreview | sean mooney proposed openstack/nova master: Remove Linux bridge VIF support from Nova https://review.opendev.org/c/openstack/nova/+/957040 | 13:23 |
| stephenfin | gibi: Could you re-apply your +2 on https://review.opendev.org/c/openstack/nova/+/986881? I'd to change two jobs to target py311 instead of py310 (see https://review.opendev.org/c/openstack/nova/+/986881/1..4) | 13:23 |
| opendevreview | Stephen Finucane proposed openstack/os-vif master: Remove linux bridge plugin https://review.opendev.org/c/openstack/os-vif/+/941586 | 13:28 |
| opendevreview | Stephen Finucane proposed openstack/os-vif master: Remove linux bridge plugin https://review.opendev.org/c/openstack/os-vif/+/941586 | 13:32 |
| gibi | stephenfin: done | 13:42 |
| stephenfin | gibi++ ty 🙏 | 13:42 |
| tafkamax | hi nova gurus. I am reading about live migration. Would you recommend setting libvirt.live_migration_permit_auto_converge to true or live_migration_permit_post_copy | 13:46 |
| tafkamax | Both are not silver bullets, but i would like to know you guyses experiences in the wild. | 13:46 |
| tafkamax | we have not gotten live migration timeouts | 13:46 |
| tafkamax | but some of them take a long time | 13:46 |
| sean-k-mooney | tafkamax: i generally recoemment prefering post copy over auto converge | 13:53 |
| sean-k-mooney | if you enable both that is what libvirt will do internally | 13:53 |
| tafkamax | ok, i read that in the docs that post-copy seems to be preferred | 13:54 |
| sean-k-mooney | tafkamax: if you have a new enough nova you can also enabel multile tcp connection for live migration which is you pair that with a high bandwith nic helps alot | 13:54 |
| sean-k-mooney | they are technially for 2 diffent things | 13:54 |
| sean-k-mooney | post-copy is for memory bound vms which is typiclly the main issue | 13:54 |
| sean-k-mooney | auto converge is a diffent aprpoch to the same problem where the vm is paused for longer and longer downtimes | 13:55 |
| sean-k-mooney | basiclly to prevent ti form dirting the memory in the first place | 13:55 |
| sean-k-mooney | but post-copy has much lower guest performance impact | 13:55 |
| tafkamax | ok thanks for the info. will look into the tcp connection thing aswell | 13:56 |
| sean-k-mooney | the tadeoff is that it extend the tiem where if there is a network interuption the guest can crash. | 13:56 |
| sean-k-mooney | that is a very very rare event | 13:56 |
| sean-k-mooney | but that is the reason its not on by default upstream | 13:56 |
| sean-k-mooney | we have had post copy on by defualt in our downstream disto for a few years now | 13:57 |
| opendevreview | Merged openstack/nova master: Drop Python 3.10 https://review.opendev.org/c/openstack/nova/+/986881 | 16:01 |
| melwitt | elodilles: hey I saw that you had WIP to fix grenade-skip-level-always on stable/2025.1 but I noticed now the failure looks more fundamental i.e. "AttributeError: module 'setuptools.build_meta' has no attribute 'prepare_metadata_for_build_editable'. Did you mean: 'prepare_metadata_for_build_wheel'?" | 18:20 |
| melwitt | not sure if there is a way to fix this or if this is a situation where the job would be better off removed from the branch. FYI for your interest :) | 18:21 |
| opendevreview | Dan Smith proposed openstack/nova-specs master: Add unpin-az spec https://review.opendev.org/c/openstack/nova-specs/+/986539 | 18:22 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!