From: ric@Opus1.COM
Date: Fri Feb 10 2006 - 11:26:35 EST
I have a group of NFS clients (Solaris 8) that have been talking to my
Linux NFS server with no problems for many months. One of the Solaris 8
systems (an UltraSparc 60) was unused, so I wiped it clean, and installed
Solaris 9 (09/05 media)+recommended patch cluster. Once I added the 20
or so user accounts that NFS mount their home directories, and started
a cron job that looks in ~/.plan, ~/.project, and ~/.signature for each
user to supplement its word list for checking for really bad passwords,
the Solaris 9 system began Panicing about 3 nights out of 7.
I rebuilt the same box with Solaris 8 (02/02 media)+recommended patch
cluster and the Ultra-60 is stable as a rock again.
Our local net is busy around midnight when this job runs, so all of the NFS
clients grumble "NFS server homer not responding" followed by "NFS server
homer ok" occasionally. The Solaris 9 client adds to this things like
WARNING: NFS server homer returned a bad file type for root
panic[cpu0]/thread=300018a8ac0:
BAD TRAP: type=34 rp=2a1005709d0 addr=28c407a82b mmu_fsr=0
or
BAD TRAP: type=31 rp=2a1003d1550 addr=10 mmu_fsr=0 occurred in module \
"unix" due to a NULL pointer dereference
Has anyone seen Solaris 9 NFS clients falling over in cases where Solaris 8
clients don't? I've pawed thru Sun's patch database without getting
anywhere.
Thanks,
Ric Anderson (ric@opus1.com)
_______________________________________________
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 - 23:38:56 EDT