[Bucardo-general] seems I found a bug...

[Michelle Sullivan]

Greg Sabino Mullane wrote:
> There is some new code in git if you want to give it a shot 
> or eyeball it. Look for "GMT" :)
>
>   

Eyeballed - it should work - will test later, however just as a note -
it will not help (infact it will stop) any existing entries in the
_delta table that are yet to be synced... So for me I'd have to restart,
or stop the input to the DB before doing the update.

Regards,

Shells

PS: my problems are back with bucardo_delta and bucardo_track..
currently it's been updating bucardo_track for 14 hours.. this is on a
brand new 24 core box with 16G RAM and a 20 spindle top spec hardware
RAID10 set.  Think this is the make_delta inserts going in... the
replication was super fast (60 minutes to read and transfer over an
Au-USA link, 10 minutes to write the changes to the DB) for the 1m rows
it just updated, but it's still going for the make_delta :-/