Bodhi 5 leaking connections #8466

New issue

Closed

opened 2019-12-15 03:14:54 +00:00 by pbrobinson · 10 comments

pbrobinson commented

2019-12-15 03:14:54 +00:00

While working from Australia bodhi is almost unusable via cli and the web interface pretty much times out every time.

$ fedpkg update
Could not execute update: Could not generate update request: ServerError(https://bodhi.fedoraproject.org/updates/, 504, Error returned from json module while processing b'https://bodhi.fedoraproject.org/updates/': b'Expecting value: line 1 column 1 (char 0)'
b"<html>504 Gateway Time-out\nThe server didn't respond in time.\n</html>\n")
A copy of the filled in template is saved as bodhi.template.last

While working from Australia bodhi is almost unusable via cli and the web interface pretty much times out every time. <pre> $ fedpkg update Could not execute update: Could not generate update request: ServerError(https://bodhi.fedoraproject.org/updates/, 504, Error returned from json module while processing b'https://bodhi.fedoraproject.org/updates/': b'Expecting value: line 1 column 1 (char 0)' b"<html><body><h1>504 Gateway Time-out</h1>\nThe server didn't respond in time.\n</body></html>\n") A copy of the filled in template is saved as bodhi.template.last </pre>

pbrobinson commented

2019-12-15 03:16:56 +00:00

Author

a 'fedpkg update' will often require 2-3+ attempts to get an update, the web interface to edit an update for adding bugs, or updating the metadata description basically times out and doesn't update it but doesn't provide a useful error either.

kevin commented

2019-12-15 03:22:10 +00:00

Can you try again now? This might be an issue we have been trying to fix all week. ;(

I just restarted the bodhi web pods... which usually clears up the issue (for a while).

Can you try again now? This might be an issue we have been trying to fix all week. ;( I just restarted the bodhi web pods... which usually clears up the issue (for a while).

pbrobinson commented

2019-12-15 03:25:44 +00:00

Author

That did work, I've been trying for 4 hours to get that submitted. It was extremely frustrating :-/

smooge commented

2019-12-15 17:06:12 +00:00

So the issue isn't with latency or you being in Australia. The issue is that bodhi, openshift, or the DB is leaking database connections and so goes into longer and longer la-la land. [All examinations of the items seem to point to the next thing in the stack as being the problem so it is probably a little of all 3.]

smooge commented

2019-12-16 18:34:18 +00:00

Metadata Update from @smooge:

Issue priority set to: Waiting on Assignee (was: Needs Review)
Issue tagged with: bodhi, database

**Metadata Update from @smooge**: - Issue priority set to: Waiting on Assignee (was: Needs Review) - Issue tagged with: bodhi, database

cverna commented

2019-12-19 10:02:23 +00:00

Metadata Update from @cverna:

Issue assigned to cverna

**Metadata Update from @cverna**: - Issue assigned to cverna

kevin commented

2020-01-06 17:41:04 +00:00

Can you try again now? This may have been related to all our koji/bodhi/rabbitmq issues and should be over now...

cverna commented

2020-01-07 19:28:01 +00:00

As @abompard mentioned there is a high possibility that the problem we have experienced with Bodhi was related to #8477.

I propose to close this ticket since Bodhi is working as expected now.

As @abompard mentioned there is a high possibility that the problem we have experienced with Bodhi was related to #8477. I propose to close this ticket since Bodhi is working as expected now.

kevin commented

2020-01-07 19:33:54 +00:00