-
-
Notifications
You must be signed in to change notification settings - Fork 2.1k
Synchrotron stopped relaying new messages #7345
Comments
Restarting the worker seemed to get it running again. |
Ah, and it quickly fell apart. The master process appears to be screaming a lot about:
|
Oh, it's worth pointing out that I might not have been quite up to date...I'll try forcing myself to latest develop and see if that helps. |
it did not |
Yup there are still bugs here as per #7308. Half-Shot can you share a stacktrace with latest develop? |
|
sorry @Half-Shot can you include the ERROR line above the traceback? |
(this looks closely related to #7308 but a slightly different code path) |
@richvdh my logs have since rotated but the traceback was cased by the error below it (it was looping the same error and traceback in the logs over and over). I can try to run develop again and capture fresh logs, if useful. |
ah don't worry. Will have a stare at the code and try and track it down. |
I'm struggling to figure out which stream could be throwing that assertion. I'll add some better logging and then get you to try again. |
Oh wait, it says:
but then... that makes no sense, because as of #7337, the events stream doesn't go via |
fresh logs from develop |
thanks, looks like a dup of #7308 |
I noticed today that my messages had stopped arriving after some time, but /sync was happily running as normal (no clientside errors). I checked the synchrotron and noticed that I was hitting the long poll timeout (30s) for my user pretty much reliably, and the sync token wasn't advancing.
And to be sure, I checked that my federation workers were all working, and they seemed to be processing incoming traffic just fine. I've also checked my certificates, and they are okay too.
It has been going like this from 23:38 last night up until now.
The text was updated successfully, but these errors were encountered: