[Bucardo-general] Swap replication latency

[Alex Balashov]

Any idea why it would kill the database in the process of ballooning out this far?  When I logged in I found 100+ postmaster connection processes and a load average of ~2.50+ and a very unresponsive database (until Bucardo was stopped).  This was the "source" database in my de facto master->slave setup that's really a swap.

--
Alex Balashov - Principal
Evariste Systems LLC
260 Peachtree Street NW
Suite 2200
Atlanta, GA 30303
Tel: +1-678-954-0670
Fax: +1-404-961-1892
Web: http://www.evaristesys.com/

On May 4, 2011, at 10:59 AM, Joshua Tolley <josh at endpoint.com> wrote:

> On Wed, May 04, 2011 at 10:46:22AM -0400, Alex Balashov wrote:
>> I've had to shut down our replication indefinitely because the 
>> bucardo_delta table got to 13.3 million rows, and none of them were 
>> being cleared out even with a purge interval of 5 minutes.  I assume 
>> that's because these were unprocessed deltas.
>> 
>> The "run time" of the last good sync ballooned out to 45 hours.  This 
>> is for a database of about ~1 GB with only 650,000 inserts, 3,510 
>> updates and 81 deletes.
> 
> Yep, based on the "last good" value, certainly Bucardo was having problems
> processing something. It most likely tried to synchronize, failed, tried
> again, failed again, and so on. Generally you'll find the reason in bucardo's
> logs, preceded by "Warning".
> 
> -- 
> Josh Tolley   josh at endpoint.com   801-987-0252
> End Point Corporation  http://www.endpoint.com