From: Ian Dickens (ian@mitre.org)
Date: Tue Aug 08 2006 - 10:00:32 EDT
Good morning,
I am seeing an issue that has me stumped and am hoping someone else
out there has seen (or better still - conquered) this issue. I have
a Storedge 3310 and Ultra320 LVD/SE card connected to a Sun Fire 440
running Solaris 8 patched with a patch cluster of approx 10-12 months
ago. The s3310 is configured for RAID0 (per project request) and has
a valid volume on it (configured by attaching to a Sun Fire V240
running Solaris 10). What is happening is that we cannot see the
s3310 correctly - format always shows it as an unknown disk type and
the error "mode sense reports nsect value as 255, adjusting to -63".
After some searching I was able to find some email traffic pointing
to some specific patches (which we installed on the system [some
where present]) with no positive effects. I am now attempting to
download the latest firmware for the s3310 but Sun's site seems
pegged at the moment.
We originally had this s3310 RAID system and Ultra320 card attached
to a Sun Fire V240 running Solaris 10 but noticed response issues
when writing large files from clients (all Solaris 8 with the same
patch set [10-12 months ago]) which is causing an Oracle database
install to fail waiting for io. Basically, on machines all running
at 1000Mbps-full when writing a large file over NFS to the volume on
the s3310 one cannot do a simple "ls -l" on the NFS volume until the
write has completed. I have been reading that perhaps increasing the
nfsd arguements from "-a 16" to "-a 256" would increase performance,
but I am not if that would help the problems on the client side.
Ideally, we would love to keep the Storedge 3310 on the Sun Fire 440
running Solaris 8 for performance reasons.
Any help more than welcome,
Ian
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:40:33 EDT