I have had this issue a couple of times now. on Solaris, and on 7.3 of the db.
I was did a normal vacuum and then I did a full which hung on a table, I waited for over an hour but there were no activity going on so I ctrl'cd the job in a terminal window, kill off all associated processes and try to do another normal vacuum which hung at the same spot. So I once again killed the processes. I went into the db and query the pg_tables tables and found the order of the tables, I vacuumed the ones just before and just after and they went fine but the offending table would not vacuum, Next I thought I could just pg_dump that table but the db would not allow me to do that either, I found a copy of the table from a restore I did on another box( the table has only 7 rows.) So I try to drop it but that would work either. So the only alternative was to stop and restart the db which is totally unacceptable in a 24x7 shop.
Does any one else see what I have done wrong or could do differently next time. Oh yes and after the stop and start I could do a full vacuum on that table as well as the whole database.
I know it's after st. Patrick's day but I did not find any green men lurking around to be causing this, any ideas?, hints?, suggestions?
I am at my wits ends as to where to turn next, I most certainly do not want to another vacuum full.
Mark M. Huber
3035 E. Patrick Lane
Las Vegas, NV 89120
<<Mark M. Huber.vcf>>
pgsql-admin by date
|Next:||From: Richard Dockery||Date: 2004-03-19 18:37:50|
|Subject: Dynamic IP authentication|
|Previous:||From: Litao Wu||Date: 2004-03-19 17:41:59|
|Subject: postgresql.conf and pg_ctl reload|