500 Internal Server Error (Ticket 0 does not exist.) when submitting some new tickets

[andrewdied@…]

Sometimes on my 0.10.2 system I get a 500 Internal Server Error when submitting a ticket. The ticket does get added. I have one custom field defined that I have left blank. I use tracd with these options:
tracd —port 8000 -r -d —pidfile=/usr/local/trac/trac.pid —basic-auth bugs,/usr/local/trac/bugs.htpass,bugs.com /usr/local/trac/bugs

This is a Solaris 10 Sparc (220R) with python 2.5, postgreSQL 8.1.3, Opera 9.02 on winxp sp2 for a browser.

The only things I see in the log are

2006-11-16 15:53:40,499 Trac[notification] DEBUG: Sending SMTP notification to tango.example.com on port 25 to [u'tickets@example.com']
2006-11-16 15:53:41,242 Trac[main] WARNING: 500 Internal Server Error (500 Internal Server Error (Ticket 0 does not exist.))

Normally I leave assign to blank and let it use the component owner as the default. I have received the error when I assign the owner explicitly, however. I do use restrict_owner = true. I have not been able to identify any particular data that cause the error, and there is nothing in the postgres logs except autovacuum.

[andrewdied@…]

This may be a dup of #4152. That one was written on 0.10, this is 0.10.2, same system. I only have 32 tickets in my system overall, so I don't have a lot of running data on it yet. Feel free to close out the appropriate ticket.

[andrewdied@…]

Jonas, I see you are the owner for this and #4152. Is there any more information I can get you on this? I still get this error pretty regularly.

[cboos]

Does this still happen?
Even with 0.10.3?

[andrewdied@…]

The Solaris box this was running is gone, and before I had the chance to upgrade it to 0.10.3 from 0.10.2. I migrated the system to a suse 10.2/trac 0.10.3 linux box, and haven't seen it since. I don't know what the fix was, but I haven't seen the error since.

Closing as worksforme, since I can't even reproduce it. (Trying to submit again, since the first time I got hit by a bayes filter.)