G'day ...
I've got a script that runs on all the servers that dump's IP traffic
data to a 7.4 database ... they all run at the same time, but I'm starting
to get the following on a reasonably regular basis:
ERROR: deadlock detected at /usr/local/abin/ipaudit2ams.pl line 175.
The code that is causing it, from the times I've been able to catch it,
is a simple update to the same table:
$upd->execute( $traffic{$company_id}{$ip_id}{$port}, $company_id, $ip_id, $date ) || die $upd->errstr;
Now, the scripts are wrap'd in a BEGIN/END ... if a file fails to be
loaded, I want the whole thing to rollback ... the deadlock itself, I'm
presuming, is because two servers are trying to update the same
$ip_id/$port/$company_id record, at the same time ...
Now, reading the DEADLOCKS section at:
http://www.postgresql.org/docs/7.4/static/explicit-locking.html
This is to be expected ... but, other then breaking the transaction
itself into smaller chunks, or staggering the scripts run times ... is
there something I'm overlooking to eliminate this? I could increase the
deadlock timeout, as an option as well ...
----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664