> Hi
> 
> I am running an utility, composite of many shell
> scripts, where a HP-UX machine is remotely executing
> a shell script in Solaris 10 machine. That script
> calls other scripts in the same Solaris 10 machine
> and takes 10-12 hours to complete. But we are facing
> random behavior. Sometimes the SunOs scripts are
> failing at different places (like echo command, mv
> command etc - rally awkward !) or sometimes the SunOs
> scripts are running, but the HP-UX process has died.
> This problem has never happened in Solaris 9 OS.
> 
> We have tried with ssh instead of remsh, but the same
> problem was noticed.
> 
> Anybody else has faced the same problem ?

Is the current Solaris 10 system located somewhere a bit different
than the old Solaris 9 system was?  If so, and if the script produces no
output for long periods of time, maybe you are going through
a stateful firewall that you weren't before, and it has limited state capacity
and times out apparently inactive connections to reclaim state table entries.
(I've seen _that_ happen quite often...)

I think processes on the remote side might well die with SIGHUP if the
connection is dropped, which would _not_ produce core dumps (but
could be detected by process accounting, for example).
 
 
This message posted from opensolaris.org
_______________________________________________
opensolaris-help mailing list
[email protected]

Reply via email to