RE: RAC 'gc blocks lost' in stress test
- From: Yong Huang <yong321@xxxxxxxxx>
- To: oracle-l@xxxxxxxxxxxxx
- Date: Wed, 19 Mar 2008 14:05:58 -0700 (PDT)
I posted this message earlier about the increase of 'gc blocks lost' in the
stress test of our 10.2.0.3 RAC (Linux x86_64). A few days later, "ifconfig
bond1" no longer showed significant packet drop (Did the OS and network team do
something?), but Oracle's 'gc blocks lost' continued to climb in the test.
This morning, we upgraded Oracle to 10.2.0.4. The problem went away! The total
value of this stat is only 6 across all 8 nodes during our 2-hour test, even
though the number of active non-idle sessions is even higher than before,
between 300 and 500 (select count(*) from gv$session where wait_class != 'Idle'
and status = 'ACTIVE'). I checked 10.2.0.4 List of Bug Fixes (Note:401436.1)
but didn't find anything obviously relevant.
Yong Huang
____________________________________________________________________________________
Be a better friend, newshound, and
know-it-all with Yahoo! Mobile. Try it now.
http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
--
http://www.freelists.org/webpage/oracle-l
- Follow-Ups:
- RMAN clone of 9i DataGuard standby instance
- From: Crisler, Jon
Other related posts:
- » RAC 'gc blocks lost' in stress test
- » RE: RAC 'gc blocks lost' in stress test
- » RE: RAC 'gc blocks lost' in stress test
- » RE: RAC 'gc blocks lost' in stress test
- » Re: RAC 'gc blocks lost' in stress test
- RMAN clone of 9i DataGuard standby instance
- From: Crisler, Jon