Re: Excessive "latch: row cache objects" waits with simple INSERT statement
Date: Mon, 5 Oct 2020 12:10:55 +0200
Message-ID: <a67b4c79-f3aa-0509-6430-444b6c909832_at_mgm-tp.com>
The trace files reveal some more weird behaviour.
For many SELECT statements, it seems that Oracle gathers statistics dynamically even though the relevant table have up-to-date statistics.
And the query itself (which does not use the GTT) causes "latch: row cache objects" as well.
I am not sure if 10KB email body would be accepted by the mailing list, so I have pasted the relevant part of the trace file here: https://pastebin.com/A26Z4MHn
(I obfuscated the table and user names in there)
Thomas
> Have I said anything yet about creating a logon trigger (on one of
> your schemas) to enable SQL tracing with waits events enabled for
> just the SQL_ID of the critical statement - perhaps with a limited
> lifetime so you only capture a few occurrences of the statement.
-- http://www.freelists.org/webpage/oracle-lReceived on Mon Oct 05 2020 - 12:10:55 CEST