Re: downgrades in v$sql_monitor
Date: Wed, 10 Jul 2013 13:25:30 +0200
Message-ID: <51DD44AA.5080203_at_www.sqltools-plusplus.org>
Hi Josh,
it might depend on how your other queries work so that they indicate a downgrade.
Remember that you might see a lower *activity* than expected in case of Parallel Execution skew where only a part of the assigned Parallel Execution Servers actually do the work (data distribution skew), or only a limited number of them is active concurrently (temporal skew).
Depending on what you try to measure this *might* look like a downgrade but actually is a skew problem.
If I ever find the time to put together the planned video tutorial on that topic I could simply point you to that - but it's not available yet unfortunately.
Randolf
Am 09.07.2013 18:07, schrieb Josh Collier:
> Hi,
>
> I set up a job to capture the contents of v$sql_monitor once a minute
> during my ETL run last night. I am having a hard time writing a query to
> find evidence of queries downgraded to serial. My other queries indicate
> that there were operations downgraded, but I cannot find them in my
> version of V$sql_monitor. All the px_servers_requested contents have the
> same value in px_server_allocated.
>
> Thanks for your time,
>
> Josh C.
>
-- http://www.freelists.org/webpage/oracle-lReceived on Wed Jul 10 2013 - 13:25:30 CEST