reosarevok: it's trying to connect to floyd for some reason
2023-10-23 29631, 2023
atj
even though the config contains no references to floyd
2023-10-23 29606, 2023
human joined the channel
2023-10-23 29608, 2023
human has quit
2023-10-23 29651, 2023
atj
reosarevok: i think this is related to the postgres changes bitmap made yesterday. there is no pgbouncer container running on floyd, but critiquebranz is trying to connect to one
2023-10-23 29631, 2023
reosarevok
Can we bounce them elsewhere? :P
2023-10-23 29617, 2023
human joined the channel
2023-10-23 29642, 2023
aaditya joined the channel
2023-10-23 29601, 2023
aaditya has quit
2023-10-23 29646, 2023
human has quit
2023-10-23 29626, 2023
human joined the channel
2023-10-23 29602, 2023
mayhem
safe travels home akshaaatt -- you're the last person to return from the summit. next time you're in barcelona we'll unravel the dark mysteries of how shower curtains work. another exciting trip, for sure!
yeah, we should depricate that piece, and just tell people to use LB API, but I'll got restart it.
2023-10-23 29631, 2023
atj
reosarevok: I could probably hard code some IPs, but I don't really want to mess with stuff I don't fully understand, especially when it's database related
2023-10-23 29622, 2023
reosarevok
Seems ok to wait for bitmap
2023-10-23 29642, 2023
bitmap
mayhem fixed it already but basically you just need to restart the container if that happens
2023-10-23 29615, 2023
mayhem
critiquebrainz-web-prod is what I restarted
2023-10-23 29602, 2023
bitmap
if we have any other projcets using uwsgi it'd be nice to compare their setup and find out why cb's is borked
2023-10-23 29652, 2023
atj
bitmap: ffs :P
2023-10-23 29610, 2023
atj
i restarted uwsgi and it was still broken, so i thought a container restart was pointless
2023-10-23 29643, 2023
bitmap
yeah! I tried restarting the runit service at first too but it wouldn't take
2023-10-23 29615, 2023
atj
I guess I should trust in the old three-finger salute method more
2023-10-23 29651, 2023
bitmap
aerozol: thanks for the posts, was gonna ping you today but should've done it sooner :)
2023-10-23 29654, 2023
bitmap
yvanzo: zas: hi, so in order to move sir-prod off of floyd and onto hendrix, I understand we have to whitelist hendrix's IP somewhere. could you help with that?
one is for master/primary connections, one for slave/standby connections, and another for "any" (unused, but useful if we want to distribute read only queries to either the master or standby in the future)
2023-10-23 29651, 2023
atj
would be good to work out what this uwsgi issue is
2023-10-23 29654, 2023
zas
bitmap: then we should adapt so we can get stats for all pgbouncer instances
2023-10-23 29611, 2023
reosarevok
It says "One limitation of the approach used here is that it doesn't detect relationship changes." - maybe you have any suggestions for that :)
2023-10-23 29626, 2023
bitmap
reosarevok: not yet, but I can check it later today
2023-10-23 29630, 2023
reosarevok
Perfect
2023-10-23 29641, 2023
reosarevok
Would be good to put a beta out tonight or tomorrow
2023-10-23 29606, 2023
reosarevok
But I see you're quite busy today :D
2023-10-23 29650, 2023
bitmap
zas: the relevant containers now are postgres-jimmy, postgres-hendrix, pgbouncer-master, pgbouncer-slave, pgbouncer-any, and haproxy-postgres
2023-10-23 29621, 2023
zas
bitmap: ok, thanks, I'll add them
2023-10-23 29658, 2023
bitmap
if there is an easy way to add haproxy stats to telegraf, it might be useful too
2023-10-23 29611, 2023
bitmap
pgbouncer connects to haproxy, haproxy to pg
2023-10-23 29643, 2023
zas
bitmap: yes, I was thinking about it, we may need to make few changes to haproxy configs so telegraf can gather stats from them, I'll have a look at this this week
2023-10-23 29652, 2023
zas
bitmap: we need to have /stats endpoint accessible on haproxy containers on hendrix/jimmy, it runs on port 8182 according to config, but this port isn't exposed by containers it seems
I had a great meeting with monkey where we discussed some of the blockers for this PR.
2023-10-23 29648, 2023
reosarevok
Other than that I spent time working on getting spark to work with the new filters, and fixing some frontend tests.
2023-10-23 29649, 2023
reosarevok
Thats it for me
2023-10-23 29650, 2023
reosarevok
'''
2023-10-23 29638, 2023
reosarevok
yvanzo said:
2023-10-23 29639, 2023
reosarevok
'''
2023-10-23 29647, 2023
reosarevok
Last week I worked mostly on actionable tasks following the summit: writing notes, prepping weblate, fixing small issues, triaging tickets, and a bit of support…
2023-10-23 29650, 2023
reosarevok
'''
2023-10-23 29617, 2023
reosarevok
aerozol said:
2023-10-23 29618, 2023
reosarevok
'''
2023-10-23 29623, 2023
reosarevok
It’s a holiday/long weekend here, and Lucy-Belle’s birthday, so I almost forgot it was Monday! I didn’t though. Lucky lucky you.
2023-10-23 29635, 2023
mayhem
stitches? what did I miss?
2023-10-23 29606, 2023
mayhem
sounds like an unusal merge conflict
2023-10-23 29610, 2023
reosarevok
aerozol said:
2023-10-23 29611, 2023
reosarevok
'''
2023-10-23 29613, 2023
reosarevok
It’s a holiday/long weekend here, and Lucy-Belle’s birthday, so I almost forgot it was Monday! I didn’t though. Lucky lucky you.
2023-10-23 29616, 2023
mayhem
oy, not too bad I hope!
2023-10-23 29617, 2023
reosarevok
Last week I mainly worked on other client work, but I did a bit more work on the summit 23 blog post. I will post it Wednesday.
Other than that I made some changes to the Fresh Releases redesign mockup, that Ansh has made great progress on, and noodled about with some design considerations with monkey
2023-10-23 29645, 2023
reosarevok
I also added YouTube videos/walkthroughs to a bunch of wiki pages
2023-10-23 29646, 2023
reosarevok
The person who is second (after me) most likely to forget it’s Monday can go next
2023-10-23 29648, 2023
reosarevok
'''
2023-10-23 29603, 2023
monkey
Mon-what?
2023-10-23 29604, 2023
reosarevok
I'm not sure how likely FishFish is to forget it's Monday but here's his review:
