| *** suzhengwei has joined #openstack-masakari | 05:50 | |
| yoctozepto | morning | 06:01 |
|---|---|---|
| yoctozepto | #startmeeting masakari | 06:01 |
| openstack | Meeting started Tue May 11 06:01:05 2021 UTC and is due to finish in 60 minutes. The chair is yoctozepto. Information about MeetBot at http://wiki.debian.org/MeetBot. | 06:01 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 06:01 |
| *** openstack changes topic to " (Meeting topic: masakari)" | 06:01 | |
| openstack | The meeting name has been set to 'masakari' | 06:01 |
| yoctozepto | #topic Roll-call | 06:01 |
| *** openstack changes topic to "Roll-call (Meeting topic: masakari)" | 06:01 | |
| yoctozepto | \o/ | 06:01 |
| suzhengwei | o/ | 06:01 |
| yoctozepto | #topic Agenda | 06:03 |
| *** openstack changes topic to "Agenda (Meeting topic: masakari)" | 06:03 | |
| yoctozepto | * Roll-call | 06:03 |
| yoctozepto | * Agenda | 06:03 |
| yoctozepto | * Announcements | 06:03 |
| yoctozepto | * Review action items from the last meeting | 06:03 |
| yoctozepto | * CI status | 06:03 |
| yoctozepto | * Backports pending reviews | 06:03 |
| yoctozepto | * Monitors docs proposals https://review.opendev.org/q/topic:add-docs+project:openstack/masakari-monitors | 06:03 |
| yoctozepto | * Xena planning -> https://etherpad.opendev.org/p/masakari-xena-ptg | 06:03 |
| yoctozepto | ** Deprecate process monitor? | 06:03 |
| yoctozepto | * Open discussion | 06:03 |
| jopdorp | o/ | 06:03 |
| yoctozepto | #topic Announcements | 06:06 |
| *** openstack changes topic to "Announcements (Meeting topic: masakari)" | 06:06 | |
| yoctozepto | I don't have any today | 06:06 |
| yoctozepto | you? | 06:06 |
| suzhengwei | none | 06:06 |
| jopdorp | None | 06:07 |
| yoctozepto | #topic Review action items from the last meeting | 06:07 |
| *** openstack changes topic to "Review action items from the last meeting (Meeting topic: masakari)" | 06:07 | |
| yoctozepto | I think none, but let's check | 06:08 |
| yoctozepto | none | 06:08 |
| yoctozepto | #topic CI status | 06:08 |
| *** openstack changes topic to "CI status (Meeting topic: masakari)" | 06:08 | |
| yoctozepto | looks green | 06:08 |
| yoctozepto | though while we are at it | 06:08 |
| yoctozepto | #link https://review.opendev.org/c/openstack/masakari/+/790216 | 06:09 |
| yoctozepto | this needs merging soon as openstack is moving forward with sqlalchemy version | 06:09 |
| yoctozepto | we seem to be well-covered except for the unit tests that get fixed there | 06:09 |
| yoctozepto | ekhm, it was meant to encourage you to review it :D | 06:13 |
| yoctozepto | anyways, let's move on | 06:13 |
| yoctozepto | #topic Backports pending reviews | 06:13 |
| *** openstack changes topic to "Backports pending reviews (Meeting topic: masakari)" | 06:13 | |
| yoctozepto | no backports atm | 06:13 |
| yoctozepto | #topic Monitors docs proposals https://review.opendev.org/q/topic:add-docs+project:openstack/masakari-monitors | 06:13 |
| *** openstack changes topic to "Monitors docs proposals https://review.opendev.org/q/topic:add-docs+project:openstack/masakari-monitors (Meeting topic: masakari)" | 06:13 | |
| yoctozepto | so, it seems the lack of documentation for monitors is a pain point for our users | 06:14 |
| yoctozepto | we have these 3 changes proposed | 06:14 |
| yoctozepto | one is much smaller and does a very minimum job to get sample config rendered (and needs some fixing still) | 06:15 |
| yoctozepto | though thanks to noonedeadpunk for reraising this docs issue | 06:15 |
| yoctozepto | I suggest we review the other 2 thoroughly and decide on the content that can go in right away | 06:16 |
| yoctozepto | or even merge the ideas from them | 06:16 |
| yoctozepto | as they conflict at the moment | 06:16 |
| yoctozepto | your thoughts? | 06:17 |
| yoctozepto | guessing none | 06:20 |
| yoctozepto | #topic Xena planning -> https://etherpad.opendev.org/p/masakari-xena-ptg | 06:20 |
| *** openstack changes topic to "Xena planning -> https://etherpad.opendev.org/p/masakari-xena-ptg (Meeting topic: masakari)" | 06:20 | |
| yoctozepto | Deprecate process monitor? | 06:20 |
| yoctozepto | do we know of a legit use case of process monitor that actually rescues instances (the goal of masakari)? | 06:21 |
| *** vishalmanchanda has joined #openstack-masakari | 06:22 | |
| yoctozepto | are you still with me? :-) | 06:23 |
| jopdorp | What does it do exactly now | 06:23 |
| jopdorp | And why is that not needed anymore? | 06:23 |
| yoctozepto | the 2nd question is another take at my question | 06:24 |
| yoctozepto | and the answer to 1st is | 06:24 |
| yoctozepto | it monitors selected process on the local system (usually nova-compute, libvirtd) and triggers actions to restore them | 06:25 |
| yoctozepto | it also notifies the masakari api at some point | 06:25 |
| yoctozepto | and the masakari engine disables the nova compute service iirc | 06:25 |
| suzhengwei | Think about taht the process runs in container. | 06:25 |
| yoctozepto | yeah, in a containerised environment the process monitor is currently a no-go | 06:26 |
| jopdorp | I think that is a good reason | 06:27 |
| yoctozepto | also, for the restoration of process one relies on systemd in the native case or docker (or anything else) in the containerised case | 06:27 |
| yoctozepto | jopdorp: which and good reason for what? | 06:27 |
| jopdorp | For deprecating | 06:28 |
| yoctozepto | ah, understood | 06:28 |
| suzhengwei | Some users include me, run their openstack cloud services based on kubernetes. | 06:29 |
| yoctozepto | so I understand you don't use it, jopdorp, because you rely on kolla ansible and we don't support it ;-) | 06:29 |
| jopdorp | Indeed | 06:29 |
| yoctozepto | suzhengwei: oh, nice! I will try to remember that! so - are you using OpenStack Helm? and you *don't* use process monitor, right? | 06:30 |
| suzhengwei | yes. | 06:31 |
| yoctozepto | yes to both questions? :-) | 06:32 |
| suzhengwei | we use helm to install openstack, no process-monitor. | 06:32 |
| yoctozepto | thanks | 06:33 |
| yoctozepto | I will check process monitor related code once again | 06:33 |
| yoctozepto | and write to the mailing list about deprecation | 06:33 |
| suzhengwei | We've been doing this for three months. | 06:33 |
| yoctozepto | also, I would love to get insight from noonedeadpunk re process-monitor (and how it is in OSA) | 06:34 |
| yoctozepto | suzhengwei: oh, nice; what were you using previously? how has been the journey with kubernetes so far? | 06:34 |
| suzhengwei | kolla-ansible | 06:34 |
| yoctozepto | oh, sorry to see you go ;-( | 06:35 |
| suzhengwei | Masakari runs ok. | 06:36 |
| yoctozepto | yeah, it does | 06:36 |
| yoctozepto | #topic Open discussion | 06:37 |
| *** openstack changes topic to "Open discussion (Meeting topic: masakari)" | 06:37 | |
| yoctozepto | anything you want to discuss? | 06:37 |
| jopdorp | Nothing for me | 06:38 |
| suzhengwei | Nothing from me. | 06:38 |
| yoctozepto | all right, thank you for attending | 06:40 |
| yoctozepto | #endmeeting | 06:40 |
| *** openstack changes topic to "This channel is *LOGGED* | Whiteboard: https://etherpad.opendev.org/p/masakari-whiteboard" | 06:40 | |
| openstack | Meeting ended Tue May 11 06:40:51 2021 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 06:40 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/masakari/2021/masakari.2021-05-11-06.01.html | 06:40 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/masakari/2021/masakari.2021-05-11-06.01.txt | 06:40 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/masakari/2021/masakari.2021-05-11-06.01.log.html | 06:40 |
| jopdorp | Thanks! | 06:42 |
| noonedeadpunk | yoctozepto: sorry, just got online | 07:19 |
| noonedeadpunk | eventually regarding masakari-monitors in osa we had bug quite till now regarding the way we configure them, which has been covered with https://review.opendev.org/c/openstack/openstack-ansible-os_masakari/+/789243 | 07:22 |
| noonedeadpunk | eventually that's why I went through reading code regarding config options and raised doc topic | 07:22 |
| yoctozepto | ack, np | 07:25 |
| yoctozepto | I was also asking about process monitor | 07:25 |
| yoctozepto | we want to deprecate it as no core seems to be using it and its value is discussable | 07:26 |
| noonedeadpunk | but except | 07:36 |
| noonedeadpunk | (disregard, wrong window) | 07:36 |
| noonedeadpunk | regarding process monitor not really so sure. Eventually it helps to restart things if it detects that they are down and not by service status, like I'd do with internal monitoring, but with presence of process... | 07:38 |
| noonedeadpunk | while I agree it's really discussable, but it's also kind of handy.... | 07:38 |
| noonedeadpunk | and disabling compute service when libvirt can't be started or not running is kind of helpful | 07:39 |
| yoctozepto | thanks for your pov, I will consider this (also for docs :-) ) | 07:45 |
| *** mgoddard has quit IRC | 07:45 | |
| noonedeadpunk | I mean I think it kind of matters what it takes to maintain that monitor. In case it's failing and hard to keep it alive, then probably worth dropping it and focus on other things. | 07:49 |
| *** suzhengwei has quit IRC | 07:50 | |
| *** mgoddard has joined #openstack-masakari | 07:50 | |
| yoctozepto | well, the issue is it is not really tested in CI so we might just go best effort here and keep it until it starts getting really problematic | 07:52 |
| yoctozepto | otoh, this is a outside-world-kind of problem | 07:52 |
| yoctozepto | as in, we are providing this as part of masakari so should in general understand the benefits of it | 07:53 |
| *** alexe9191 has joined #openstack-masakari | 08:11 | |
| alexe9191 | Good day everyone:) | 08:12 |
| alexe9191 | I had a question about masakari. Does it need a shared storage? | 08:12 |
| alexe9191 | And does the instances live or cold migrate? | 08:12 |
| yoctozepto | hi alexe9191 | 08:14 |
| alexe9191 | hey yoctozepto | 08:15 |
| yoctozepto | it requires shared storage for instances to migrate when their current host is down | 08:15 |
| yoctozepto | the instances evacuate (so it's like they migrate cold) | 08:15 |
| yoctozepto | the host goes down -> the instances are evacuated | 08:16 |
| alexe9191 | What if the host does not go down? can masakari monitor nova queue for instance for errors or instances for error and cold migrate them somewhere else? | 08:17 |
| yoctozepto | ok, then we have an instance monitor that can detect that an instance went down and trigger its restart (it's up to nova to move it) | 08:25 |
| yoctozepto | and it will not move it if the original host is healthy | 08:25 |
| alexe9191 | Nice. This is exactly what I was looking for. | 08:42 |
| alexe9191 | Thank you:) will test it out | 08:42 |
| yoctozepto | you are welcome | 09:00 |
| *** alexe9191 has quit IRC | 09:35 | |
| yoctozepto | noonedeadpunk: continuing our morning discussion on the usefulness of processmonitor - nova-compute will detect libvirt not responding (being down or just unavailable on the selected address) and report being down - therefore I can't see how processmonitor helps it | 17:45 |
| yoctozepto | (it looks like I'm trying to drop it at all costs but in all honesty I am looking to find ways to tell people where processmonitor really is useful) | 17:45 |
| *** vishalmanchanda has quit IRC | 18:34 | |
| *** yoctozepto has quit IRC | 19:13 | |
| *** yoctozepto6 has joined #openstack-masakari | 19:13 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!