[Pgpool-general] pgpool-II dies when running a 'vacuum analyze'

[Bruno Lustosa]

Hello, list.

I'm running pgpool-II as a load balancer with 2 backends.
Recently, I deleted a huge table from the database (around 36 million
rows), and tried to run a vacuum analyze after that to reclaim the 15
or so gigabytes it occupied.
The vacuum runs for some time, and then the pooler just dies. On the
client, I get the message about 'backend closing connection
unexpectedly'.
The backends are running postgresql 8.2.4, and pgpool-II is 2.0.1,
fetched from CVS on Feb 20, I think.
Although I could run the vacuums separately on each backend, it would
be nice to be able to run it directly from the pooler. Our maintenance
procedure that runs nightly deletes lots of things, and we run it on
the pooler to avoid the (rare, now) case when one of the backends are
offline. So, the ideal place to run the vacuum would be after the
maintenance.
How can I provide more details? The log file from the pooler?

Thanks!

-- 
Bruno Lustosa <bruno at lustosa.net>
http://www.lustosa.net/