[Bucardo-general] Missing records,

Kaveh Mousavi Zamani kaveh at kavehmz.info
Thu Apr 14 01:48:22 UTC 2011

Checking delta and track tables is a great idea. But unfortunate for me, we
run "bucardo.bucardo_purge_delta" and "bucardo.bucardo_purge_q_table" every
month and track and delta tables are clean of any useful information.

But I checked the postgres logs,

All 7 missing records, are from a period of time (in the same minute) that
slave DB had many problems (connection maxed, out of memory, several heavy
and amateurish SQL queries killing the server, and at the end slave was
killed,I think by users command). But I am not sure if theoretically this
can be cause of problem.

In a very simple way, is this the flow of process in bucardo?

Begin Master Transaction
Begin Slave Transaction
    Read From Master
    Update Salve
    Update Master
Commit Slave
Commit Master

On Wed, Apr 13, 2011 at 11:08 PM, Greg Sabino Mullane <greg at endpoint.com>wrote:

> ...
> > 7 in around 60M records in one of table, and some more here and there in
> > other tables.
> >
> > Is it possible that bucardo failure to connect to slave be cause of these
> > missing reocrds?
> Very unlikely. The delta rows are not removed until we've committed them
> to the slave. Can you check if the bucardo_delta table on the master
> still has the rows, and if there is a corresponding bucardo_track entry?
> --
> Greg Sabino Mullane greg at endpoint.com
> End Point Corporation
> PGP Key: 0x14964AC8
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mail.endcrypt.com/pipermail/bucardo-general/attachments/20110414/1609bd99/attachment.html 

More information about the Bucardo-general mailing list