Librarian: psycopg2.OperationalError: could not send data to server: Connection timed out
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Launchpad itself |
Triaged
|
Critical
|
Unassigned | ||
Storm |
Fix Released
|
Low
|
Sidnei da Silva |
Bug Description
smagoun1 began getting intermittent 'Processing failed' errors trying to load private attachments from the librarian e.g.
https:/
Found these erros in the librarian logs:
2012-08-09 21:05:40+0000 [-] Unhandled error in Deferred:
2012-08-09 21:05:40+0000 [-] Unhandled Error
Traceback (most recent call last):
File "/usr/lib/
File "/usr/lib/
File "/usr/lib/
--- <exception caught here> ---
File "/srv/launchpad
result = context.call(ctx, function, *args, **kwargs)
File "/srv/launchpad
return self.currentCon
File "/srv/launchpad
return func(*args,**kw)
File "/srv/launchpad
return func(*args, **kwargs)
File "/srv/launchpad
return func(*args, **kwargs)
File "/srv/launchpad
ret = func(*args, **kwargs)
File "/srv/launchpad
alias = self.storage.
File "/srv/launchpad
return self.library.
File "/srv/launchpad
File "/srv/launchpad
result = self._store.
File "/srv/launchpad
return Connection.
File "/srv/launchpad
File "/srv/launchpad
return Connection.
File "/srv/launchpad
File "/srv/launchpad
return function(*args, **kwargs)
lifeless was able to reproduce and began stress testing
We ruled out pgbouncer and postgres
After restarting the librarians the error did not return even under stress test.
Related branches
- Free Ekanayaka (community): Approve
- Jamu Kakar (community): Approve
-
Diff: 56 lines (+21/-2)2 files modifiedstorm/databases/postgres.py (+1/-0)
tests/databases/postgres.py (+20/-2)
tags: | added: canonical-losa-lp |
Changed in storm: | |
status: | Fix Committed → Fix Released |
milestone: | none → 0.21 |
I suspect that we suffer a network 'glitch' for some other unidentified reason. Then the connection is bad, and we're not actively closing it, rather trying to reset it, or some such thing.