Re: ASSM in 10g RAC doesnt seem work that well

Christo:

I've lost track of what's going on.

I'm not sure that you're both testing the same thing. I think there may be one set of results that is for a RAC 2-node cluster, and another that is non-RAC. It also seems possible that whilst your CPU results are session based, the other set are system-based, which could make a lot of difference when you consider the effects of global cache collisions on the processes handling GES and GCS

Going back to your original figures (or the things I think were your original figures), the CPU time difference for the two sessions as you switched from ASSM to non-ASSM seem to be small (2%), and the major time difference (30%) seemed to be waits for global cache coherency - and it's always possible that extra CPU would be used when global cache collisions start to appear. It might be worth looking at your original tests again, and trying to get stats for the relevant background processes too.

Regards

Jonathan Lewis

http://www.jlcomp.demon.co.uk/faq/ind_faq.html The Co-operative Oracle Users' FAQ

http://www.jlcomp.demon.co.uk/seminar.html Public Appearances - schedule updated March 9th 2005

--
http://www.freelists.org/webpage/oracle-l