| *** tosky has quit IRC | 00:41 | |
| openstackgerrit | Lingxian Kong proposed openstack/trove master: Pass availability zone through to volume creation. https://review.opendev.org/761250 | 00:52 |
|---|---|---|
| *** spatel has joined #openstack-trove | 00:54 | |
| openstackgerrit | Sam Morrison proposed openstack/trove stable/victoria: Pass availability zone through to volume creation. https://review.opendev.org/761847 | 00:58 |
| openstackgerrit | Merged openstack/trove master: Fix edit instance action return code https://review.opendev.org/761694 | 01:07 |
| *** spatel has quit IRC | 01:23 | |
| *** sapd1 has joined #openstack-trove | 01:28 | |
| *** sapd1 has quit IRC | 02:18 | |
| *** e0ne has joined #openstack-trove | 02:38 | |
| *** e0ne has quit IRC | 02:43 | |
| sorrison | Hi lxkong, so when a user does an instance upgrade now the version of the DB is upgraded but the underlying image and the version of trove-guestagent stay the same? | 03:38 |
| *** e0ne has joined #openstack-trove | 04:53 | |
| *** e0ne has quit IRC | 04:58 | |
| lxkong | sorrison: yes, the instance upgrade only means the db service upgrade, not os neither guest agent. | 06:20 |
| lxkong | sorrison: trove has never provided os and guest agnet upgrade function before | 06:21 |
| lxkong | until 'rebuild' function was implemented in Victoria. | 06:21 |
| *** rcernin has quit IRC | 06:48 | |
| *** rcernin has joined #openstack-trove | 06:49 | |
| *** sapd1 has joined #openstack-trove | 07:19 | |
| *** rcernin has quit IRC | 07:38 | |
| sorrison | Well it essentially has with the instance upgrade | 07:38 |
| sorrison | I think the change in how upgrade works needs to be given more thought. | 07:39 |
| sorrison | The way upgrade has worked up until ussuri is better I think as it allows operators to push out new versions of the DB, the operating system and the trove guest agent and importantly allow the user to do that at a time that suites them | 07:40 |
| sorrison | this is a killer feature in trove. I don't see the new instance rebuild method being that useful | 07:41 |
| sorrison | We have run trove since liberty and now on ussuri and we haven't once had an outage on users trove instances that wasn't initiated by the user themselves by doing an upgrade | 07:42 |
| sorrison | We send them a notification and alert them they need to do the upgrade by a certain time (usually a few months) else we will do it for them | 07:42 |
| sorrison | This is extremely useful | 07:42 |
| *** e0ne has joined #openstack-trove | 08:02 | |
| *** e0ne has quit IRC | 08:07 | |
| *** sapd1 has quit IRC | 08:33 | |
| *** e0ne has joined #openstack-trove | 10:33 | |
| *** sapd1 has joined #openstack-trove | 10:34 | |
| *** tosky has joined #openstack-trove | 11:09 | |
| *** sapd1 has quit IRC | 11:12 | |
| *** sapd1 has joined #openstack-trove | 11:32 | |
| *** sapd1 has quit IRC | 11:46 | |
| *** spatel has joined #openstack-trove | 12:55 | |
| *** spatel has quit IRC | 12:59 | |
| *** sapd1 has joined #openstack-trove | 14:35 | |
| *** congnt96 has joined #openstack-trove | 14:47 | |
| *** e0ne has quit IRC | 18:23 | |
| *** tosky has quit IRC | 20:17 | |
| *** e0ne has joined #openstack-trove | 20:23 | |
| *** tosky has joined #openstack-trove | 20:28 | |
| *** congnt96 has quit IRC | 21:05 | |
| *** sapd1 has quit IRC | 21:05 | |
| *** e0ne has quit IRC | 21:06 | |
| *** rcernin has joined #openstack-trove | 21:27 | |
| *** e0ne has joined #openstack-trove | 21:50 | |
| *** rcernin has quit IRC | 22:06 | |
| *** e0ne has quit IRC | 22:10 | |
| *** rcernin has joined #openstack-trove | 22:12 | |
| *** rcernin has quit IRC | 22:22 | |
| *** rcernin has joined #openstack-trove | 22:22 | |
| lxkong | The OS and trove-guestagent upgrade process should be transparent to the end users. You can do that before because when upgrading database you need to re-build the guest image. | 22:49 |
| lxkong | After using docker image, it's easier to upgrade datastore for the user, doesn't involve cloud operator any more. | 22:49 |
| lxkong | It just becomes the cloud operator's responsibility to upgrade OS and trove-guestagent (by using rebuild API). | 22:50 |
| lxkong | which is mroe reasonable | 22:50 |
| openstackgerrit | Lingxian Kong proposed openstack/trove stable/victoria: Fix upgrading instance datastore version https://review.opendev.org/762011 | 23:09 |
| openstackgerrit | Lingxian Kong proposed openstack/trove stable/victoria: Pass availability zone through to volume creation. https://review.opendev.org/761847 | 23:19 |
| sorrison | yeah that is bad, we don't want to have to upgrade all out users DBs ourselves, puts more work on us and also means users can't do it when it suites them | 23:21 |
| *** tosky has quit IRC | 23:33 | |
| *** e0ne has joined #openstack-trove | 23:38 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!