*** dangtrinhnt has joined #openstack-fenix | 01:18 | |
*** hyunsikyang has quit IRC | 01:56 | |
*** hyunsikyang has joined #openstack-fenix | 03:04 | |
tojuvone | well, seems the nova conf does not offer anything useful. Looking if SIGHUP for scheduler updated the cache | 09:18 |
---|---|---|
tojuvone | by fast, this seems to do the trick and it is fast | 10:16 |
tojuvone | generic solution might not be that straight forward as currently I need to ssh from container to host to issue 'sudo docker kill -s SIGHUP nova_scheduler' | 10:21 |
tojuvone | There are bugs reported to nova with similar problems when migration simultaneously instances belonging to same antiaffinity group | 10:22 |
tojuvone | That hist Fenix, when this kind of optimiztion will be done | 10:23 |
tojuvone | This current upgrade case would not easily happen in real world, as host upgrade would take enough time for cache to be correct | 10:24 |
tojuvone | Not to mention the host should mostly be even restarted or at least the services. Bug is only when simulating upgrade without any real action on host | 10:37 |
*** tojuvone has quit IRC | 10:47 | |
*** tojuvone has joined #openstack-fenix | 10:47 | |
*** dasp has quit IRC | 21:52 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!