Conference noted::pwv50ift

4190.0. "PWRK$LMMCP in RWAST on shutdown?!!" by CHOWDA::GLICKMAN (writing from Newport,RI) Tue Mar 04 1997 16:07

At the site I support they went to shutdown the PATHWORKS server 5.0D ECO3
on OpenVMS 6.1 VAX to avoid the consumption of nonpaged pool and had a 
problem on one of the 3 nodes in this VAXcluster running PATHWORKS when 
shutting down.  These nodes are VAXA, VAXB, and VAXD with a cluster
alias of VAXNPT.

One analyst started to shutdown PATHWORKS via SYSMAN and hadn't set the
timeout value large enough (10 seconds) so the SYS$STARTUP:PWRK$SHUTDOWN
started on each node and then timed out.

The log of this looked something like this:

SYSMAN> SET ENV/NO=(VAXA,VAXB,VAXD)
%SYSMAN-I-ENV, current command environment:
        Individual nodes: VAXA, VAXB, VAXD
        Username SYSTEM        will be used on nonlocal nodes

SYSMAN> SET TIMEOUT 0:00:10
%SYSMAN-I-TIMEVAL, timeout value is 00:00:10:00
SYSMAN> DO @SYS$STARTUP:PWRK$SHUTDOWN
%SYSMAN-I-OUTPUT, command execution on node VAXA
Shutting down the currently runnings server(s)...
%SYSMAN-I-NODERR, error returned from node VAXA
-SMI-E-TIMEOUT, remote operation has timed out
%SYSMAN-I-OUTPUT, command execution on node VAXB
Shutting down the currently runnings server(s)...
%SYSMAN-I-NODERR, error returned from node VAXB
-SMI-E-TIMEOUT, remote operation has timed out

%SYSMAN-I-OUTPUT, command execution on node VAXD
Shutting down the currently runnings server(s)...
Stopping PWRK$LMMCP with STOP/ID
SYSMAN>

The analyst tells me that VAXD shutdown successfully at this point.

Another analyst shutdown PATHWORKS on VAXA with the SYS$STARTUP:PWRK$SHUTDOWN
command from the DCL prompt on VAXA and that then fine.  

The first analyst shutdown PATHWORKS on VAXB with the SYS$STARTUP:PWRK$SHUTDOWN
command from the DCL prompt on VAXB on a second terminal session from his
PC and this is the message he got: 

Stopping PWRK$LMMCP with STOP/ID
%PWRK-W_OPINCOMPL, unable to shut down all PATHWORKS processes

PWSHOW revealed that PWRK$LMMCP was in an RWAST state on VAXB at 14:18
not 14:26 (when the nonpaged pool problem was indicated to occur).

The PWRK$LOGS:PWRK$LMMCP*_NODE.LOG was checked for any errors and there
were none since the last re-boot.

One of the analysts tried a STOP/PROCESS of the PWRK$LMMCP and that did
not work.

Since after reviewing the RWAST Troubleshooting document we knew that a
re-boot was the only option, management here wanted to get PATHWORKS going
again so a re-boot was done (no crash dump was collected due to time 
constraints).

PATHWORKS was shutdown successfull on the other two nodes.  Startup
was done and looked ok but when PC clients attempted connections they
failed. Once VAXB was shutdown and the other two nodes had PATHWORKS 
restarted then PATHWORKS connections were functional from PC clients.

Again, Management here would like to know what caused PATHWORKS not to 
shutdown successfully on VAXB and why VAXB affected the other two nodes
in the VAXcluster.

Appreciating any responses.  If I need to provide more information,
let me know.

Thanks.