06:02:21 <yoctozepto> #startmeeting masakari
06:02:21 <opendevmeet> Meeting started Tue Aug  3 06:02:21 2021 UTC and is due to finish in 60 minutes.  The chair is yoctozepto. Information about MeetBot at http://wiki.debian.org/MeetBot.
06:02:21 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
06:02:21 <opendevmeet> The meeting name has been set to 'masakari'
06:02:26 <yoctozepto> o/
06:02:33 <jopdorp> o/
06:03:59 <shenxinxin> o/
06:05:53 <suzhengwei> o/
06:06:58 <yoctozepto> ok, let's start
06:07:01 <yoctozepto> with the agenda
06:07:18 <yoctozepto> * CI status
06:07:18 <yoctozepto> * Important pending reviews (important bugfixes, backports)
06:07:19 <yoctozepto> * Next release planning
06:07:19 <yoctozepto> * Open discussion
06:07:23 <yoctozepto> #topic CI status
06:08:05 <yoctozepto> all systems green :-)
06:09:28 <yoctozepto> #topic Important pending reviews (important bugfixes, backports)
06:10:00 <yoctozepto> still https://review.opendev.org/c/openstack/masakari/+/796732
06:10:44 <yoctozepto> please remember to review patches :-)
06:11:24 <yoctozepto> #topic Next release planning
06:11:31 <yoctozepto> I have seen movements in the specs
06:11:36 <yoctozepto> but I did not have time to review them again
06:11:47 <yoctozepto> though I am happy we are progressing
06:13:59 <suzhengwei> me too.
06:16:25 <yoctozepto> should we discuss anything today?
06:16:33 <yoctozepto> I will be reviewing specs this week again
06:16:50 <yoctozepto> I have the evacuations and consul monitoring in my plans
06:17:10 <suzhengwei> yes
06:18:28 <suzhengwei> I reviewed some patchs about 'add-docs' https://review.opendev.org/q/topic:%2522add-docs%2522+(status:open)
06:19:01 <yoctozepto> ah, yeah; have you incorporated the other patches in the one at the top?
06:19:27 <yoctozepto> ah, no, I see the change
06:20:05 <suzhengwei> I reviewed them without comments.
06:20:36 <suzhengwei> I think they are ok to be merged.
06:21:05 <yoctozepto> yeah, you applied the comments; thanks; will check and yeah, likely merge
06:21:55 <suzhengwei> from my side, I would like to add hostmonitor by consul based on them.
06:22:24 <yoctozepto> makes sense; we should definitely start with it documented :-)
06:23:44 <suzhengwei> I will try to update them. If you feel it is ok, please merge them.
06:24:20 <yoctozepto> ok, I will mark them for a review today
06:24:23 <yoctozepto> thanks
06:24:28 <shenxinxin> Could we talk about this spec? spec for vm evacuations for host recovery | https://review.opendev.org/c/openstack/masakari-specs/+/789432
06:25:23 <yoctozepto> ok
06:26:04 <shenxinxin> I still don't think it is necessary to include migration_uuid. We already have source_host and dest_host.
06:26:31 <shenxinxin> The migration_uuid cannot be obtained directly.
06:28:04 <yoctozepto> I mean, we don't have to include it upfront; it could be useful for audit purposes; we can include it once we make nova able to return it
06:28:58 <yoctozepto> regarding the source host - it is stored at the notification level (all instances from one notification will have the same source host - so it is cleaner not to repeat this information)
06:29:12 <suzhengwei> I don't think it is necessary to wait for nova api changing, because I don't know how long to pause.
06:29:14 <yoctozepto> we can have a very efficient join in SQL
06:29:46 <yoctozepto> suzhengwei: yes, we will not wait for nova; you can skip the migration_uuid for now
06:30:04 <yoctozepto> I hope the migration_uuid situation is clear now ;-)
06:32:12 <yoctozepto> so, to rephrase: ignore the migration_uuid for now, don't add it if you don't want to
06:32:18 <yoctozepto> now for the other two fields
06:32:30 <yoctozepto> source_host - I already wrote a line about it above
06:32:42 <suzhengwei> I suggest we move ahead with the spec description. Once nova support the new api, we modify our code.
06:33:49 <yoctozepto> yes
06:37:50 <yoctozepto> shenxinxin: do you agree regarding the source_host?
06:38:02 <yoctozepto> (I am missing your answer to my reasoning)
06:38:39 <suzhengwei> In our application, we just care about where the instance ha from and to. just the hostname is enough.
06:39:58 <shenxinxin> What I mean is that adding source_host is more intuitive, and more efficient when querying evacuation records.:)
06:42:13 <suzhengwei> It doesn't care about other things about the Host, likes type/reserved/control_attr.
06:43:57 <yoctozepto> yeah, it does not, but duplication is usually a bad thing
06:44:38 <yoctozepto> ok, I don't care that much, let's go with source_host, but perhaps name it "source_host_name" to be clear what it is
06:45:00 <yoctozepto> now about the dest_host - what do we need it for? to verify whether the instance has not been re-migrated?
06:45:26 <suzhengwei> yes.
06:45:30 <yoctozepto> ok
06:45:37 <yoctozepto> then the same thing applies, let's append _name
06:46:07 <suzhengwei> how about source_hostname and dest_hostname?
06:46:53 <yoctozepto> suzhengwei: fine by me; both variants look good
06:47:04 <yoctozepto> though source_host_name
06:47:09 <yoctozepto> makes it easier to create
06:47:14 <yoctozepto> source_host_whatever
06:47:21 <suzhengwei> ok
06:47:42 <shenxinxin> We need to store this instance from and to in the evacuation record, also for migrate back. for dest_host field.
06:47:45 <yoctozepto> I left two nit comments on the spec review as well
06:48:51 <yoctozepto> also look at the two comments on PS2, commit message: https://review.opendev.org/c/openstack/masakari-specs/+/789432/2
06:49:09 <shenxinxin> Ok,thanks.
06:49:17 <yoctozepto> I think those two name fixes, two nits to fix and the commit message
06:49:27 <yoctozepto> and I will be approving this
06:49:45 <yoctozepto> I saw the code has progressed as well so it might be quick to implement :-) and then we focus on consul, great
06:51:34 <shenxinxin> Ok, I will try to update it these days.
06:51:38 <yoctozepto> thanks
06:51:58 <yoctozepto> anything else about Xena today?
06:52:09 <jopdorp> not form me
06:52:30 <suzhengwei> not from me.
06:52:42 <shenxinxin> not from me.
06:53:52 <yoctozepto> ok
06:53:56 <yoctozepto> #topic Open discussion
06:54:05 <yoctozepto> anything else at all? :D
06:55:57 <yoctozepto> ok, thank you for meeting today
06:56:03 <yoctozepto> #endmeeting