I'm in a bit of a bad spot. We have a number of remotely deployed
postgres servers. We are very happy with the depenability of postgres,
but have recently run into a problem with pg_dump in SOME of the servers.
When doing a pg_dump databasename > file.dmp pg_dump eats up all the
memory, and is eventually killed because it takes
up a huge amount of memory. The dump file is always 0 bytes. It works
perfectly fine on our other server.
When we run pg_dump in verbose mode, it is trying to get the trigger
information.
Does anybody have any clue what could
be happening? Or even how I can help determine the problem? The machines
are only accessable to us by modem dialup.
Postgres Version 7.1.2 ( Yes, I know it's ancient - but again - these are
a number of servers remotely deployed... )
Redhat 7.1 ( see above ).
I know this is an old version of postgres, but I would really appreciate
any help I can get trying to track this down.
Steve
