I've got my Solaris 11 Software update Library on an NFS share.
I've noticed frequent large IOPS on that share, and can see that Ops Center is once every 10 minutes running "du" command across the library.
My guess is that the Monitoring Rules on this library for it's disk space are doing this, however they are unable to be modified so I can check -- the "Edit" and "Disable" actions are greyed out. They are active for the other Libraries -- all NFS shares from the same NAS -- but not for the Solaris 11 library.
Any ideas why I can't change the monitoring for the S11 library? Any ideas for finding more information?
I've had a similar problem with it running du's on SAN attached veritas filesystems which was causing a lot of problems.
I found the culprit was a script that the agent runs regularly. /opt/sun/n1gc/bin/vdiskadm-ofs.
I ended up commenting out the line that run's the du, and changing it to 'size=0' instead. That solved our problem for now. I've raised a ticket with oracle to investigate this as from what I can see it's only supposed to run on local ZFS filesystems. I can't see much point to it doing this at all anyway. Why not just use a 'df' or a similiar zfs function.
Thanks for pointing me at the culprit. Will raise a SR with Oracle, but at at least I can work around it for now.
Weirdly it stopped this behaviour at around 4am thismorning. It might just be the scheduling of this task and it will come back.
I still don't understand why the monitoring rules on my S11 library would be "locked".
Unfortunately, even commenting out the "du" in that script hasn't stopped the behaviour properly.
It looks right, but the process still runs and still hammers the storage device.
I will probably have to raise an SR if noone knows the answer to this one.