perfSONAR User Q&A and Other Discussion

[Mark Feit <>]

Andrej Ricnik writes:

 

I'm just wondering whether we're the only ones seeing tons of errors like:

 

2019-04-07 06:49:03 NZST [1562-1] ERROR:  canceling autovacuum task

2019-04-07 06:49:03 NZST [1562-2] CONTEXT:  automatic vacuum of table "pscheduler.public.run"

 

Autovacuum will self-abort to avoid holding locks that would get in the way of queries, so it’s not really an error per se, it’s just a notification that it happened.  The default interval on CentOS is one minute, which I’d assume (but can’t guarantee) carries over into Debian/Ubuntu.

 

The run table is the largest in the system in terms of number of rows and amount of activity, which makes it prone to bloat.  I don’t see too much of this in the test bed or on our production systems, so you’ve either got a really buy system or the defaults on your distro are different enough to cause problems.

 

If you wouldn’t mind doing me a favor, grep the word “vacuum” out of the last few days of logs and send me that plus your postgresql.conf off-list.  We can do some tuning of the autovacuum behavior on a per-table basis.  Doing it globally works but is a little heavy-handed.

 

--Mark