I have a four-node cluster of Solaris systems (all 280Rs) basically doing
NFS to several thousand SGI, IBM RS/6000 and Sun workstations. Today I installed
the latest Solaris 8 recommended patch cluster and now when I bring this
node in the cluster and start failing over services to it it will just lock
up. You can ping it but it stops serving NFS, you cannot get a respoonse
from the console. Once I was running "top" on it when it locked and the
last reported load average was 125 with kernel usage at 99.5%. I'm guessing
something in the kernel is spinning out of control. Has anyone else seen
this kind of behavior? Possibly which patch could have caused this?