2021-09-15 19:47:54,556 - irc3.wikibugs - DEBUG - Starting wikibugs... 2021-09-15 19:47:54,708 - irc3.wikibugs - DEBUG - Connected 2021-09-15 19:47:54,709 - irc3.wikibugs - DEBUG - CONNECT ping-pong () Exception ignored in: <object repr() failed> Traceback (most recent call last): File "/usr/lib/python3.5/asyncio/tasks.py", line 92, in __del__ self._loop.call_exception_handler(context) AttributeError: 'NoneType' object has no attribute 'call_exception_handler' Exception ignored in: <object repr() failed> Traceback (most recent call last): File "/usr/lib/python3.5/asyncio/tasks.py", line 92, in __del__ self._loop.call_exception_handler(context) AttributeError: 'NoneType' object has no attribute 'call_exception_handler' 2021-09-15 19:48:35,962 - irc3.wikibugs - CRITICAL - connection lost (140180063038656): None 2021-09-15 19:48:35,962 - irc3.wikibugs - CRITICAL - closing old transport (140180063038656)
Description
Description
Related Objects
Related Objects
Event Timeline
Comment Actions
Actually if I run it directly on a exec node it's fine too, but under the grid it's having issues.
Comment Actions
Mentioned in SAL (#wikimedia-cloud) [2021-09-15T20:03:02Z] <legoktm> redis2irc is running in a screen because of T291129
Comment Actions
So I tried:
- Running redis2irc manually on the exec node it was failing from - works
- Bumping grid memory to 2G - didn't work
- Running redis2irc from a different exec node, under grid - didn't work
- Running redis2irc with logging enabled manually (to see if logging was causing problems) - works
- Disabling TLS since we had issues with that in the past - didn't work
I'm totally at a loss of what's wrong. I'm running it in a screen now on the bastion just so it's back up.
Comment Actions
Based on discussion with @Bstorm I switched our REDIS_HOST config from "tools-redis" to "tools-redis.svc.eqiad.wmflabs", and now it works.
I'm not really sure that actually fixed it or we got lucky, but I'll take it.