Stelios Mavromichalis wrote
>> the load of the machine is also low (like 0.2).
Which means little if the update is waiting for a lock to be released by one
other process; which is more likely the situation (or some other concurrency
contention) especially as you said that this particular user generates
significant transaction/query volume (implied by the fact the user has the
most balance updates).
During slow-update executions you want to look at:
pg_stat_activity
pg_locks
to see what other concurrent activity is taking place.
It is doubtful that dump/restore would have any effect given that the
symptoms are sporadic and we are only talking about a select statement that
returns a single row; and an update that does not hit any indexed column and
therefore benefits from "HOT" optimization.
HTH
David J.
--
View this message in context:
http://postgresql.1045698.n5.nabble.com/Re-recently-and-selectively-slow-but-very-simple-update-query-tp5802553p5802555.html
Sent from the PostgreSQL - performance mailing list archive at Nabble.com.
