EM12c host metric explanations

From: Rich Jesse <rjoralist2_at_society.servebeer.com>
Date: Tue, 4 Jun 2013 09:54:26 -0500 (CDT)
Message-ID: <2515a5f40350356cf003798a9caf6c45.squirrel_at_society.servebeer.com>



In the Performance Home page of one 11.2.0.3 DB in EM 12.1.0.1.0 w/BP1, I see that the top graph of "Host: Runnable Processes" shows a significant portion in "Non-Database Host CPU". Since the server for this DB (AIX 5.3 TL12) only hosts 1 Production ERP DB and a few other minor Production DBs, I fired up nmon on AIX, but there's nothing really going on this server.

So I drilled down in EM, clicking on the "Non-Database Host CPU" item, which brought me to the host's Performance Summary page. There, the "Top 10 Processes" shows 8 processes that don't exist in a "ps -ef". And it's been that way for at least the last hour that I've been looking into this.

Possibly related, the "Total Processes" metric (under "Load") also shows a count of about 2700, but again "ps -ef" says that number is closer to 650. We did have a server crash a little over a month ago

I've mentioned here before that I have an ongoing issue with EM12c regularly reporting very large spikes in the Average Active Sessions for CPU Wait, but I can't figure out where that's coming from either.

Any one else seeing this? As we're ramping up for major migrations, I've got zero time to upgrade EM again, but right now I'm having more trust issues...

TIA! Rich

--
http://www.freelists.org/webpage/oracle-l
Received on Tue Jun 04 2013 - 16:54:26 CEST

Original text of this message