Details that were relevant--
a) the scu scan edt bus 3 hangs whether inside of sys_check or not.
b) The contents of bus 3 were two Exabyte tape stackers. Each takes
up two luns, one for the tape drive and one for the mover.
c) We hadn't seen this problem until we recently switched from
an alphaserver 2100 to an alphaserver 2100A.
alan_at_nabeth.cxo.dec.com
suggests there may be a conflict with the pci bridge (pci2000) which
is in a 2100A, but also points out that there could have been
problems with overall single-ended SCSI bus length that just
manifested themselves now. Also points out that the tape stacker
in question isn't a qualified DEC device.
Recourse is either to open a software service call with DEC (Compaq)
or patch sys_check so it doesn't do the scu scan edt.
Steve Timm
------------------------------------------------------------------
Steven C. Timm (630) 840-8525 timm_at_fnal.gov
http://home.fnal.gov/~timm/
Fermilab Computing Division/Operating Systems Support
Scientific Computing Support Group--Computing Farms Operations
---------- Forwarded message ----------
Date: Thu, 3 May 2001 12:36:01 -0500 (CDT)
From: Steven Timm <timm_at_fnal.gov>
To: tru64-unix-managers_at_ornl.gov
Subject: sys_check hangs on "scu scan edt bus 3"
I have now seen this particular behavior on my system three times.
This is an Alphaserver 2100A (recently acquired).
I am trying to run the sys_check utility on it. It
is currently sitting hung, trying to do the command
scu scan edt bus 3
as part of the sys_check. The ps output looks like this:
1335 console IW + 0:00.03 /usr/sbin/getty console console vt100
2248 ttyp1 U N 0:00.02 scu scan edt bus 3
2353 ttyp1 R + 0:00.02 ps
8265 ttyp1 I N 0:01.14 ksh -p /usr/sbin/sys_check - scsi none 33
-
31051 ttyp1 S 0:00.51 -csh (csh)
31233 ttyp1 I N 0:10.18 ksh -p /usr/sbin/sys_check
Bus 3 is otherwise quite functional. The drives on it
can each be accessed quite well. But the scu scan edt bus 3
will continue to keep going and going until such time
as the machine will be rebooted (which will be tomorrow). Eventually
if the process is left uncleared, other I/O also starts to hang,
forcing the machine to be rebooted.
I am running 4.0g with all the latest firmware and patches.
Steve
------------------------------------------------------------------
Steven C. Timm (630) 840-8525 timm_at_fnal.gov
http://home.fnal.gov/~timm/
Fermilab Computing Division/Operating Systems Support
Scientific Computing Support Group--Computing Farms Operations
Received on Fri May 04 2001 - 20:09:51 NZST