now my country needs me. to fight a glut. of beer. especially here in bavaria. See hetzner.
bai!
akshaaatt[m]
Drinks on me today ruaok, haha!
ruaok
you picked a good time. beer is often the cheapest drink on the menu. there is a law here that states at least ONE thing needs to be more cheap than beer. normally apple juice.
🤯
akshaaatt[m]
😂😂
bitmap
zas: that's great to hear. i'm sure it was the track paging added to large release pages, yes. I've rarely seen any timeouts in sentry since that was released, and I'm sure those drove up server load
Btw lucifer I think we should add some future plans to the readme of mb android app. Let me complete the final blog and this readme part by tomorrow.
bitmap
reosarevok: yea that's a legitimate error actually, since calling those functions back won't keep 'this'. you can resolve that by just wrapping them in arrow functions though. pushed a commit for that, feel free to squash as appropriate
BrainzGit
[bookbrainz-site] 14MonkeyDo merged pull request #684 (03series-entity…relationship-attributes): feat: set relationship attributes value in relationship-editor https://github.com/bookbrainz/bookbrainz-site/p...
ritiek has quit
bitmap
yvanzo: how have you recovered the sir-prod containers from this "RuntimeError: maximum recursion depth exceeded in cmp" error in the past?
I worked around it for now by moving the self.connection/self.channel assignments in the code. hopefully the queue starts dropping now
my working theory is that sometimes acking a message fails, so the action_wrapper decorator tries to reconnect to rabbitmq calling `connect_to_rabbitmq` method.
calling that method re-adds add the callbacks and stuff. and the new added callback then fails setting up a chain reaction.
One issue here is that we should never try to ack a message if the original channel/connection got closed. if we do it, rabbitmq will close the new connections as well.
the error message shown on the heading is misleading here because we try to access a property of the exception which does not exist, hence throwing another exception but if you see the details, the actual error message is found to be
`PreconditionFailed(406, u'PRECONDITION_FAILED - consumer ack timed out on channel 1', (0, 0), u'')`
bitmap
good point, would that apply to reject/requeue_message too?
lucifer
i'll need to check whether it applies to reject and requeue or not.
i had submitted a couple of patches to improve the exceptions messages for sir in sentry. but they are not deployed yet. i'll ask yvanzo when he is around and see if we can get those out soon.
bitmap
ok, if not ping me and I'll get them deployed
from your investigation though it sounds like we should be checking that the connection/channel are still open in either action_wrapper or callback_wrapper
lucifer
indeed, we should check for whether the connection/channel are opened or not and only retry if rabbitmq allows in that scenario.
MRiddickW joined the channel
piwu
lucifer: search for the user / link in the chat log; essentially submitted whole spotify dump via api in (iirc 5) batches of up to 10k at a time
idk I should probably move it to forum, now that other chat has flushed the topic