19:01:33 <nirik> #startmeeting Fedora Infrastructure Ops Daily Standup Meeting
19:01:33 <zodbot> Meeting started Thu Jan  9 19:01:33 2020 UTC.
19:01:33 <zodbot> This meeting is logged and archived in a public location.
19:01:33 <zodbot> The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:33 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:01:33 <zodbot> The meeting name has been set to 'fedora_infrastructure_ops_daily_standup_meeting'
19:01:33 <nirik> #chair smooge nirik relrod cverna
19:01:33 <nirik> #info meeting is 30 minutes MAX. At the end of 30, its stops
19:01:33 <nirik> #info agenda is at https://board.net/p/fedora-infra-daily
19:01:33 <zodbot> Current chairs: cverna nirik relrod smooge
19:01:33 <nirik> #topic Tickets needing review
19:01:47 <smooge> and I just got my browser working
19:02:32 <nirik> just in time. :)
19:02:43 <cverna> o/
19:02:48 <smooge> .ticket 8514
19:02:49 <zodbot> smooge: Issue #8514: Need someone to push for the new DNS change - fedora-infrastructure - Pagure.io - https://pagure.io/fedora-infrastructure/issue/8514
19:03:07 <nirik> thats done/not needed, but we need to update docs.
19:03:21 <smooge> ok will fix ticket
19:03:51 <nirik> well, can we quickly update the doc?
19:03:56 * nirik looks
19:04:08 <fm-admin> pagure.issue.edit -- smooge edited the close_status and status fields of ticket fedora-infrastructure#8514 https://pagure.io/fedora-infrastructure/issue/8514
19:04:09 <fm-admin> pagure.issue.tag.added -- smooge tagged ticket fedora-infrastructure#8514: dns https://pagure.io/fedora-infrastructure/issue/8514
19:04:10 <fm-admin> pagure.issue.edit -- smooge edited the priority fields of ticket fedora-infrastructure#8514 https://pagure.io/fedora-infrastructure/issue/8514
19:04:11 <fm-admin> pagure.issue.comment.added -- smooge commented on ticket fedora-infrastructure#8514: "Need someone to push for the new DNS change" https://pagure.io/fedora-infrastructure/issue/8514#comment-619999
19:04:50 <smooge> nirik, I have not been able to update the docs at all lately. it has been on my 'round-tuit' to figure out why
19:06:06 <bowlofeggs> i like to make DNS edits with a YOLO attitude, and then set the timeout to 86400
19:07:12 <nirik> https://pagure.io/infra-docs/pull-request/177
19:07:32 <nirik> .ticket 8516
19:07:33 <zodbot> nirik: Issue #8516: fedmsg-gateway on Python3 needs monitoring fixes - fedora-infrastructure - Pagure.io - https://pagure.io/fedora-infrastructure/issue/8516
19:07:44 <smooge> what is funny is I can merge on that repo
19:07:52 <smooge> which I just did
19:07:54 <nirik> so, sadly this will take some digging into the bowells of fedmsg. :(
19:08:12 <bowlofeggs> ewwww
19:08:17 <cverna> how easy would it be to move to fedora-messaging ?
19:08:23 <bowlofeggs> the bowlofeggs of fedmsg?
19:08:36 <smooge> s/bowlofeggs/boweloffedmsg/
19:08:43 <bowlofeggs> is fedmsg-gateway the thing that does <--> with fedora-messaging?
19:08:49 <nirik> no
19:08:54 <cverna> we should have the fedmsg messages on fedora-messaging
19:09:40 <nirik> I am trying to recall which thing gateway is.
19:09:51 <nirik> but it's interfacing with the public
19:10:24 <cverna> the goal of this monitoring is to check if an application send a message within the last X hours ?
19:10:48 <cverna> if that's the case we could replace this and use fedora-messaging I guess
19:10:56 <nirik> no, it's not
19:11:39 <cverna> ha ok :(
19:12:01 <nirik> fedmsg-gateway listens on all the proxy servers and gateways messages to/from busgateway
19:12:10 <bowlofeggs> it's challenging to know what to do about it if we don't know what it is
19:12:12 <nirik> I just can't recall if it's the rw one or the ro one
19:12:18 <bowlofeggs> oh
19:12:18 <nirik> the docs are not much help
19:12:38 <bowlofeggs> is this what i get messages from if i run fedmsg-tail in my house?
19:13:07 <nirik> might be yeah.
19:13:27 <cverna> oh important is it for it to be monitored :P ?
19:13:29 <bowlofeggs> and we want to know if it dies for external people by monitoring it on each node?
19:13:40 <bowlofeggs> yeah is this mission critical?
19:14:15 <cverna> I would say no, even more considering that we are asking people to switch to fedora-messaging
19:14:28 <cverna> this is even maybe something we could tiurn off
19:14:32 <cverna> turn*
19:14:47 <bowlofeggs> and we have a thing for fedora-messaging users to use from their homes that isn't this/
19:14:52 <nirik> yeah, so I think this is that... the outgoing fedmsgs
19:15:09 <nirik> I'm against just turning it off without notice.
19:15:26 <nirik> the monitoring was just to make sure it's running and not filling up
19:15:28 <cverna> yes of course, if we turn off we need to announce it
19:15:51 <nirik> it broke because we moved proxies to f31 and python3
19:16:10 <nirik> so before it was 'fedmsg-gateway' and now it's fedmsg-gateway-3
19:16:23 * nirik tries something dump
19:16:27 <nirik> dumb even
19:17:20 <bowlofeggs> i think we could announce that we're turning it off in n days and then do that, instead of figuring out how to monitor it. as long as we do have a good story for people at home tailing fedora-messaging messages
19:17:25 <bowlofeggs> we do have that too, right?
19:17:30 * bowlofeggs isn't actually sure we do
19:17:42 <cverna> bowlofeggs: yes we now have it
19:17:47 <bowlofeggs> excellent
19:17:48 <nirik> we do for that... but there is another case...
19:18:01 <nirik> fedmsg-trigger I don't think is covered?
19:18:08 <bowlofeggs> oh i don't know that one
19:18:09 <bowlofeggs> what is that?
19:18:10 <relrod> what all needs to be ported to python3 to get the current setup fixed? Is it just a script or two? Or all of fedmsg?
19:18:14 <cverna> bowlofeggs: hum actually not as simple as fedmsg-tail, it needs a little bit of scripting
19:18:33 <nirik> the nagios plugins are python2
19:18:50 <relrod> ok that seems easy enough to fix (without actually having seen them yet)
19:18:59 <relrod> I can look at porting
19:19:11 <smooge> relrod, a couple of nagios-plugins in Fedora Infrastructure and bugfix fedmsg-gateway-3 to open the right socket name
19:19:19 <bowlofeggs> oh i didn't know the current setup is actually broken, i thought it just needed monitoring
19:19:24 <cverna> yes but that's not a great use of the time, since we want to turn off fedmsg
19:19:26 <bowlofeggs> that makes it more complex
19:19:45 <nirik> broken?
19:20:00 <cverna> nirik: what is fedmsg-trigger ?
19:20:02 <relrod> cverna: but we haven't turned it off yet. Right now it's still a production service, so we should probably monitor it.
19:20:18 <nirik> the service should be working, the monitoring is broken
19:20:34 <nirik> cverna: it's a script/wrapper thing to listen for matching messages and run a script on them
19:20:47 <cverna> relrod: yes so the question is can we turn this service off, since we don't want people to use fedmsg anymore
19:20:50 <nirik> we touted it a while back as a way for people to act on things
19:21:13 <relrod> cverna: no I get that, but it's going to take time because as nirik said we don't just want to kill it without warning.
19:21:16 <bowlofeggs> oh, well if the service is working, could we just say we're shutting it down in xyz days?
19:21:22 <cverna> I think we should bring it up to the list and announce our intent to turn this service off and see the replies
19:21:46 <bowlofeggs> oh,but the trigger thing
19:22:01 <nirik> yeah, I am worried people are using that and will be mad...
19:22:02 <bowlofeggs> so if we ported fedmsg-trigger to fedora-messaging, could we then announce a shut down of the gateway?
19:22:08 <smooge> cverna, you will need to let people know both internal and external.
19:22:11 <relrod> I think we should handle #8516 and "do we want to shut it down?" as two separate tickets
19:22:17 <relrod> and I will claim 8516 and look at it
19:22:35 <nirik> https://github.com/fedora-infra/fedora-messaging/issues/155
19:22:38 <smooge> cverna, bowlofeggs I believe the trigger stuff is used internally for "things"
19:22:44 <fm-admin> pagure.issue.assigned.added -- codeblock assigned ticket fedora-infrastructure#8516 to codeblock https://pagure.io/fedora-infrastructure/issue/8516
19:23:00 <bowlofeggs> i guess if it's easy to fix the monitoring that could be good. but if it's not easy, and we can shut it down "soon", that could be a reasonable response too
19:23:01 <cverna> I think we will have a lot of fun post colo move :P
19:23:10 <smooge> relrod, I started looking at what was needed for porting and will pm you what I found
19:23:11 <cverna> finding out about all these things :P
19:23:15 <bowlofeggs> but i guess that also asks how easy the trigger thing is to port
19:23:23 <relrod> smooge: thanks
19:23:29 <nirik> looks like there is work already merged on the trigger?
19:23:36 <nirik> (see above RFE github link)
19:23:47 <bowlofeggs> if trigger is impportant, it should be ported anyway, because fedmsg is not reliable
19:24:49 <nirik> sure. we do want to retire fedmsg.
19:24:55 * nirik wonders how much is left...
19:25:02 <cverna> looking at the github ticket the trigger part is few line of codes
19:25:06 <bowlofeggs> prob way more than i think is left ☺
19:25:14 <bowlofeggs> because i never heard of any of this until today
19:25:18 <cverna> https://github.com/fedora-infra/fedora-messaging/issues/155#issuecomment-483252548
19:25:19 <bowlofeggs> which makes me think there's more i don't know hahah
19:25:39 <nirik> the known unknowns.
19:25:56 <nirik> ok, so relrod is taking that monitoring thing, we already have a ticket for the trigger... move on?
19:25:58 <bowlofeggs> and there can be unknown known unknowns
19:26:00 <cverna> yes there is still quite a lot of work, without talking about fedmsg-meta-.. repo and fmn and badges etc ..
19:26:05 <bowlofeggs> known unknowns that you forgot you knew
19:26:33 <nirik> so, I see off hand, that the upload.cgi isn't moved yet. ;)
19:26:41 <cverna> Do we want to gauge if we should turn the service off or not ?
19:27:03 <cverna> I am happy to send a mail to the list, to see who is using that
19:27:16 <nirik> I'd prefer to have a aswer on the trigger before we do, but I guess if thats going to be a long time...
19:27:52 <cverna> I think we have a answer, it is a script that people have to do rather than a cli
19:27:58 <smooge> cverna, the lists I believe you will need to talk to are devel, infra but I don't know what the internal ones are since I only know about their usage when it is broke
19:27:59 <cverna> is that acceptable ?
19:28:34 <cverna> smooge: yes hopefully the internal users are looking at the public list too
19:29:17 <nirik> I guess it's a solution... seems a but more complex, but should work (I haven't tested it)
19:29:24 <smooge> hahahahhahahahahahaha.. <BREATH> hahahahahhaha
19:30:15 <cverna> I guess we run out of time :)
19:30:23 <nirik> yeah...
19:30:45 <nirik> we can end, or I wanted to talk with bowlofeggs about helping with some backlog...
19:31:19 <smooge> ok lets do another meeting
19:31:21 <bowlofeggs> nirik: we can do a 1x1 if you want for that
19:31:23 <smooge> #endmeeting