Conference decwet::windows-nt

5854.0. "SWXCRMON stops logging" by UTRTSC::VISSER () Thu Mar 27 1997 13:27

Cross-Posted in SSAG::ASK_SSAG
    
It has taken us a long time to understand WHY the SWXCRMON stopped
reporting.
Now we understand HOW to reproduce it BUT we have NO idea where the REAL
problem is: SWXCRMON or WNT or the combination or the way they interact.

Anybody to give us a helping hand?
					Jan Visser.

The following issue has been observed with the KZESC/SWXCR/KZPSC Monitor
utility SWXCRMON as used with Windows NT V3.51 (both Intel as Alpha).
The KZESC is at FW-rev 2.16 or the KZPSC at rev 2.36.

The SWXCRMON is installed as a SERVICE within WNT, so the monitor starts as
soon the system boots. This results in entries in the SYSTEM EVENT LOGGING,
whenever a problem occurs with a disk behind this RAID-controller.
Additionally, it is setup within the SERVICE to "Allow service to Interact with
Desktop". This allows the SWXCR-log to be popped-up large, notifying the
local System Administrator of a problem.
SWXCRMON is V1.35.

This works fine (in general).
HOWEVER, it has been observed that sometimes (even if the status of a Logical
Drive behind the RAID-controller changes [e.g. fails, or end of rebuild])
NO entries appear in the Event Logger and NO messages are added to the SWXCR-
log file are added. The SWXCRMON-service however is STILL THERE!

WORKING SCENARIO:
	1. Boot WNT
	2. Logon as Administrator on console
	3. Remove disk from RAIDset
	4. The SWXCR-log pops-up, displaying a disk is failed
	5. Re-insert disk
	6. The SWXCR-log displays: Rebuild started
	7. At the end of the rebuild: SWXCR-log shows Rebuild completed.
All SWXCR-log events are also shown in the System Event Logger.

FAIL SCENARIO 1:
	1. Boot WNT
	2. Logon as Administrator on console
	3. Remove disk from RAIDset
	4. The SWXCR-log pops-up, displaying a disk is failed
	5. Re-insert disk
	6. The SWXCR-log displays: Rebuild started
              All SWXCR-log events are also shown in the System Event Logger
	      up to now.
	7. Logoff from console
	8. Rebuild finishes after some time
	9. Logon as Administrator again.
	10.NOW the "Rebuild Completed" is NOT present in the Event Logger, NOR
	   is it added to the SWXCR-log file.

FAIL SCENARIO 2:
	1. Boot WNT
	2. Logon as Administrator on console
	3. Double-click SWXCRMON to make it large and MINIMIZE it again.
	4. Logoff from console.
	5. Remove a disk from a RAIDset.
	6. This last Failure is NOT reported in the Event Logger, nor in the
	   SWXCRMON-LOG file and ANY futher activities are NOT logged.

The only way to recover from those FAIL Scenario's is to:
	a. Reboot     or
	b. Stop the SWXCRmon service and Start it again.

The REAL problem is for those systems, which are monitored REMOTELY, where
the Event Logger is the ONLY place to look for RAID-controller/disk problems.
If a local administrator logs in, it MAY STOP any SWXCRMON-logging!