Re: CPU usage after killing an Oracle session

mlanda_at_vnet.ibm.com writes:

>In <3sbtnr$hfe_at_post.gsfc.nasa.gov>, joanne_at_eosdev1.gsfc.nasa.gov (Joanne Woytek) writes:
>>Has anyone experienced situations where an Oracle session is cancelled,
>>either through the alter system kill session command or with a ctrl-C and
>>Oracle continues to chug along using 90% of the CPU? This has happened
>>to us a number of times now from different users under different
>>circumstances. I plan on trying some systematic testing but was
>>wondering if others had ideas as to what might be happening. We are
>>running Oracle 7.1.4 on an SGI IRIX 5.3 system.
>>
>>Thanks
>>
>>Joanne Woytek
>>Code 902.2
>>NASA/GSFC
>>joanne_at_daac.gsfc.nasa.gov
>
>
>Could these sessions be active and performing some I/O at the time the kill session
>was issued? If they are, they are considered uninteruptable (for instance,
>performing a rollback) until the operation is complete. This means all resources
>associated with this session will be reserved until the operation is complete.
>
>
>M.Landa

I have seen it on couple of other platforms. We traced the shadow process (begin tracing before killing the session, e.g. truss on SVR 4) and the problem was with I/O subsystem. If you are using some kind of Disk Volume management system etc.; it is the first thing you should look for. It gets stuck in a read() or write() call and burns CPU....check it out.

Sumant Chaudhari Received on Thu Jul 13 1995 - 00:00:00 CEST