Tom Lane wrote:
Found it --- the actual leak is in index_create, not in TRUNCATE or
CLUSTER at all; and it's been there a really long time.
Patch for 7.4 branch attached.
Excellent! Does index_create refer to something that is invoked as a
consequence of CREATE INDEX? I'm looking through
Tom Lane wrote:
TRUNCATE and CLUSTER both rebuild indexes, so they'd also trigger the
leak.
Sorry to bug you again, but I have two quick followup questions: (1) is
the leak you discovered fixed on the 8.0 branch? and (2) would closing
the database connection once per day be a reasonable way
Tom Lane wrote:
I was sort of expecting you to come back and say that you
thought the process might have done 640K TRUNCATEs over its lifespan,
but I guess not?
That's possible. The process does twelve TRUNCATEs every minute. The
problem we're talking about seems to occur only when the
Tom Lane wrote:
I suppose what we are looking at here is some operation that is
invalidating a relcache entry but failing to clear it.
Hm. After discussing this with people here we have a hypothesis. The
process that issues the TRUNCATE command does something a little
peculiar: every