[Spacewalk-list] Run away DELETE query eating my disk space

[Jon Miller]

Hello,

I have a situation where it appears that a DELETE statement is eating my disk
space via tmp table spaces and eventually fills the file system and then
rolls it back. Then starts up again and I'm observing this cycle thanks to
space alerts. By eating disk space, I'm talking about swings of close to 50G
of space that occurs in the span of 44min. 

Here is query being issued:
  DELETE FROM rhnPackageChangeLogData
    WHERE id NOT IN ( SELECT DISTINCT changelog_data_id FROM rhnPackageChangeLogRec )

I had earlier added an extra 20G lun to our Spacewalk (2.0) server but that
was before I realized what was happening. I'm curious if I can 1. determine
how much disk space I'll really need and 2. is there a better way to perform
this clean job? 

Per #2, I'm very tempted to shutdown spacewalk, make a DB backup and then
figure out a way to run that DELETE without transaction support, though I'm
not sure how to do that with PostgreSQL yet.

Before I get too crazy, I thought I'd ask folks here for advice.

Thanks,
-- 
Jon Miller