[Bucardo-general] A problem with custom conflict resolution.

[Lev Tannen]

Hello,

I recently run into a problem with bucardo conflict resolution that caused
me a day of agony. While it was pure my error, I think it is useful to share
my experience.

I have a master to master replication with a custom conflict resolution. It
worked fine for a long time so I have forgotten how I set it. I decided to
do some minor change in my Perl routine that processes a conflict. I
modified the routine and updated a record in the customcode table to reflect
my changes. The result was that all  records in the customcode_map table
disappeared?!!! I discovered this and entered a new record in the
customcode_map that implied custom conflict resolution on the sync level.
This was a mistake on my part, because the manual states that a custom
conflict resolution is always done on the goat level.
The result was an error message in my conflict log when I tried to start
bucardo. The message was:

DBI::db=HASH(0x8901f48)->disconnect invalidates 4 active statement handles
(either destroy statement handles or call finish on them before
disconnecting) at /usr/lib/perl5/site_perl/5.8.5/Bucardo.pm line 3397

*???????????????*

*I spend a day before I found the cause: a wrong record in the
customcode_map*.

Lev Tannen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mail.endcrypt.com/pipermail/bucardo-general/attachments/20091007/bcdae218/attachment.html