HELP detailed interpretation of the attached COLLECT command

[ddramos]

hi guys,

i have a GS160 machine used as sybase database server. It has 8cpu and 16gb memory. I am monitoring the system using the collect command to gather point in time cpu and memory utilizations. I am just wondering why there seems to be a lot of WAIT value for CPU on RAD 0 (the system has 2 QBB shelves having 4cpu and 8gb mem). Attached is a sample output from collect -scm. Kindly help in interpreting the file in greater detail. Thanks.

regards,
ddr

[Joris Denayer]

ddr,

Much more info is needed in order to give an answer.

OS-Version (# sizer -v)
HW configuration
Disk & Tape configuration (SAN or direct SCSI)
Network configuration
What processes are running on the QBB's
.
etc...

Don't look further, if you're still running on V4.0G. This version is not fully NUMA-aware. It was made to give the possibility to run V4 on a GS-platform.
If your application permits it, upgrade to v5.1B.

Joris

[ddramos]

hi joris,

I am running tru64 V5.1B.
This a dedicated database server to two front end application servers.
Disk access is via SAN

I'd like to know what RUNQ means from CPU Summary and the normal LOAD AVERAGE output, because as i am observing the collect info i've noticed that RUNQ is steady at 1 and the load average is ascending with values greater than 5.

regards,
ddr