"(We will migrate to whatever the latest stable version in the 12.x series is)." sounds like you couldn't be bothered to check, do you mean "at that point"?
real-only time sounds like the safest kind, but I guess you wanted read-only?
And "but that we will have some amount" I would probably drop that "that" there but I guess it works both ways
Otherwise it seems fine :)
c1e0 has quit
CatQuest
reo
Pac23
gsoc ideas up ?
yvanzo
working on it!
sumedh joined the channel
Pac23
cool,
yvanzo++
yvanzo
updating beta.mb.o
Pac23
karma bot broken ? how do i do ++ here ?
yvanzo
Pac23: there will be something about sir :)
No karma bot here, we have BrainzBot instead.
Pac23
oh cool,idk if i am capable enough of doing a complex gsoc project :/ too scared of failing gsoc.
so how do i call the brainzbot ? doing nick++ did not show a prompt here on my hexchat window.
ruaok
reosarevok: > "at that point"?
because we won't know which version in the 12.x series will have been released.
Mr_Monkey
!m Pac23
BrainzBot
You're doing good work, Pac23!
Mr_Monkey
That's how :)
Pac23
ah,thanks !m Mr_Monkey
!m Mr_Monkey
BrainzBot
You're doing good work, Mr_Monkey!
Mr_Monkey
Only works with the "!m" at the beginning of the sentence though
That :D
Pac23
!m yvanzo
BrainzBot
You're doing good work, yvanzo!
Pac23
i see,lol
ruaok
reosarevok: thanks for the tips, I made three small changes. (or not so small in the "real" case)
ruaok publishes it
Gore|woerk has quit
Matthew_ joined the channel
Matthew_ is now known as Guest76865
Guest76865
Hello! Thanks for the blog post about the Postgres upgrade. I just wanted to hurl a slight spanner from us replicated slave merchants in your direction... Neither AWS nor Azure currently provide a Postgres 12 managed database service...
ruaok
hiya!
do you have an appreciation for how far that support lags behind the PG releases?
we've got over 3 months....
and also, note that it isn't CRITICAL that you update.
it is recommended.
Guest76865
I'm not sure. Not a _huge_ amount. I can do some research and ask about AWS about their roadmap for that. I've no contacts at Microsoft.
ruaok
what if we refrained for (AWS|Azure) 12.x support + 30 days before we use 12.x features?
I have some contacts, but I have a feeling i know what the response will be. (read: not forthcoming)
yvanzo
updating prod mb.o
ruaok
thanks yvanzo.
Guest76865
Sure. I understand. I'm not too concerned about it in the short term and they will eventually make the version available. Let me see if I can get someone in the know to give an idea... Anyway, just wanted to give you a heads up!
ruaok
thanks for that, much appreciated.
and like I said, if we need to hold off using PG 12.x features, we can do that.
but for our own sanity, we really want to use this cadence cycle for moving to 12.x
ok, great. let's touch base closer to the time, shall we?
Guest76865
Sure. ruaok - while I'm here. We've been running the SIR for a while and, aside from the artist_credit ID 1 issue, it's been largely fine. However, we have noticed some of the indexing processes hanging every now and then, requiring a kill / restart. Is that something you've seen at all?
Etua joined the channel
ruaok
yvanzo: zas: any thoughts to contribute to this? ^^
Guest76865
Currently, we are restarting SIR before fetching replication packets and it seems to be working fine.
ruaok
seems less than ideal.
zas
more yvanzo than me, I know nothing about SIR stuff
Guest76865
The code is quite.... cryptic. For me at least!
ruaok
well, its not SIR, but more kicking-SIR. :)
Guest76865
It's possibly environmental - resource exhaustion (not enough provisioned) but I just wondered if you'd seen similar
ruaok
I think what we've seen is also caused by resource exhaustion. adding more resources seems to have greatly improved the stability of search.
Guest76865
OK. Thanks ruaok.
ruaok
np. thanks for your quick response on this.
Guest76865
np
yvanzo
I confirm the resource exhaustion, but it is still unclear how to address it, provisioning more isn't necessarily enough.
It is more an issue of tuning configuration, but that is sometimes difficult to adjust.
Guest76865
Yes yvanzo, it's quite possibly the number of configured processes. I suspect the part of the issue is in the loop logic as the only way to stop a process once it gets stuck is a `kill -9`