217359 - Mid-air collision caused 500 Server Internal Error

Status: RESOLVED WORKSFORME

(mozilla.org Graveyard :: Server Operations, task)

Description

[Olivier Cahagne]

How it happened:
1. Entered comment in bug 217296, submitted it,
2. In the mean time, bug 212796 comment 11 was submitted by someone else,
3. 500 Internal Server Error,
4. Shift-Reload bug 212769, submitted comment,
5. bug 212769 comment 12 successfully submitted.

Didn't try to reproduce. Didn't find dupes.

It's the first time I get a 500 Internal Error on bugzilla, so it could be
related to the machine switch.
I should have got a Mid-air collision.

According to http://bugzilla.mozilla.org/show_activity.cgi?id=217296, no other
between both comments.

Comment 1

[Olivier Cahagne]

sorry for the spam, every previous reference to a bug report was about bug 217296.

Comment 2

[Myk Melez [:myk] [@mykmelez]]

I've moved the master Bugzilla database to the same machine as Bugzilla.  Before
Bugzilla was accessing the master database via an SSH tunnel across the
Internet.  I'm still seeing "Aborted connection" errors in the MySQL error log,
although I'm no longer seeing 500 errors in the Apache error log, so this may be
fixed.  If you see this error again, please report the URL it happened on, the
approximate time it happened, and your IP address if available so I can track it
down in the log.

Comment 3

[Olivier Cahagne]

I hope you can still get this info in the log:
 - URL: http://bugzilla.mozilla.org/show_bug.cgi?id=217296 (actually, rather,
the form being POSTed -> process_bug.cgi ),
 - Time: 2003-08-26 11:23 or 11:24,
 - IP@: 193.251.x.x or 193.252.x.x I cannot remember exactly, but this is
usually the range of IP@ I get from my ISP.

Comment 4

[Myk Melez [:myk] [@mykmelez]]

11:23 in the evening or the morning?

Comment 5

[Olivier Cahagne]

that must be morning (was evening for me, 9h time difference I believe), and I
actually took 11:23 as per displayed by bugzilla (bug 217296 comment 12).

Comment 6

[Dave Miller [:justdave]]

Which would have placed the timing for your error before the final switch was
thrown to move it all to the new server.

Myk, have there been any other reports since the DB master was switched?  If
not, I guess I would write this off to version compatibilities from a MySQL 4.0
database acting as a replication slave to a 3.23 database.

Comment 7

[Dave Miller [:justdave]]

*** Bug 217381 has been marked as a duplicate of this bug. ***

Comment 8

[Myk Melez [:myk] [@mykmelez]]

It could also have been the SSH tunnel.  I haven't heard any reports of problems
since we moved over to the 4.0 master database which doesn't use an SSH tunnel,
so this problem seems to be solved by that change.  Please reopen if you see the
problem again.