[torqueusers] torque 6.1 mom getting into bad state

[John Griffin-Wiesner]

Since updating to torque 6.1 we've had the moms on our compute
nodes frequently getting into a bad state. The nodes will show
up as down in torque. There will be a load of 1 even though
nothing is going on, and there will be some job files from
finished jobs left in /var/spool/torque/aux and
/var/spool/torque/mom_priv/jobs. The mom daemon will have to be
killed with -9.

A regular 'service pbs_mom stop' sometimes gives:
kERROR: cannot shutdown mom daemon on localhost (errno=97-Address family not supported by protocol)

Anyone else seen this?