[Bucardo-general] Kid Dies When using Syslog for Logging

[Greg Sabino Mullane]

> Well I think I might of fixed the kid death issue. I've upgraded to Perl
> 5.8.9 (from 5.8.8), reinstalled all the perl modules I had installed,
> upgraded  plperl to 8.2.11, and upgraded to PostgreSQL 8.2.11 (from
> 8.2.9) server (client was already at the version). I have syslog and
> debugfile turned on, and nothing has died in the last 2 and half hours.
> I've been doing huge deletes of 9000+ records at a time, as well as
> inserts of similar size. I'm guessing this may be able to be racked up
> to a bug in Perl itself or one of the modules (when I was doing the
> upgrades no Perl modules changed versions, just reinstalled). However
> since I went for the "upgrade everything on the system" approach I don't
> know exactly what change made it work. Any guesses?

Not really, but I hope it was a non-Bucardo bug, cause I can't make
heads or tails of it otherwise. :)

> I'll be doing more extensive testing tomorrow to see if I can get it to
> die.

Cool. log_line_prefix, please!


> Lastly, the bug when bucardo_ctl is run without --dbhost still exists.
> Any steps you'd like me to do to track this down?

I can't really think of anything except to check the logs and see if you
can determine about when this happens in the code. Maybe adding some
purposeful die calls will help narrow it down?

-- 
Greg Sabino Mullane greg at endpoint.com
End Point Corporation
PGP Key: 0x14964AC8

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 226 bytes
Desc: OpenPGP digital signature
Url : https://mail.endcrypt.com/pipermail/bucardo-general/attachments/20090123/bb02110c/attachment.bin