This is an rx4640 running 11.31 and Oracle 10g.
We've run into a big user-interaction problem. It appears that anything that requires I/O to vg00 is causing a wait. It does not seem to affect Oracle. It's running on a SAN presented volume in a different volume group.
This problem comes and goes, its very frustrating.
There are no hardware errors in the chassis log. Nothing unusual gets written to the syslog. The only thing I can find that might be a clue is an entry at the bottom of dmesg:
"Warning: The validity of the tunable values could not be completely verified, because the value of the tunable 'process_id_max' will not be known until the system is booted. The tunable values will be verified during boot. Please check the console messages during boot to see if there are any tunable value errors."
I checked process_id_max in kcweb, it was set to 30000. I upped it to 65535. Problem persists.
I've checked all the logical volumes on vg00, all the extents are current and vgsync completed with no errors on it, so I do not think that I have a bad disk at this time.
this is a production system, the earliest I can reboot it would be 4am tomorrow morning. Anything I can look at before then?