Server Crashed

[dpellegrini]

Anyone have any ideas on this one?

Yesterday the Groupwise server locked up. It appears as though it was the result of someone sending a HUGE e-mail attachment. The server we rebooted, downed gracefully and rebooted again. Then during the volume load, there was an error message:
GWINTER: Invalid cache path specified under WebPublisher Settings.
GWINTER: Terminating WebAccess Server

1-10-01 6:39:05 pm: SERVER-4.11-2715
Module did not release 14016 resources
Module: GroupWise Web Access Server
Resource: Low-level Allocation Memory
Description: Alloc Memory (Bytes)
Then the server ABENDED(the abend below):

Server SEMBLER_MAIL halted Wednesday, January 10, 2001 6:01:21 pm
Break 1: Server-4.11a: Page Fault Processor Exception (Error code 00000000)

Registers:
CS = 0008 DS = 0010 ES = 0010 FS = 0010 GS = 0010 SS = 0010
EAX = E0238B40 EBX = E0238B40 ECX = 00000118 EDX = 00000000
ESI = E0238B40 EDI = 00000000 EBP = 1D27AC08 ESP = 1D27AC08
EIP = F11430CC FLAGS = 00017282
F11430CC 8B58FC MOV EBX,[EAX-04]= ?
EIP in THREADS.NLM at code start +0000A0CCh

Running process: gwinter 0 Process
Created by: GWINTER.NLM
Stack pointer: 1D27AC58
Stack limit: 1D25B010
Scheduling priority: 0
Wait state: 00
Stack: --1D27AC24 ?
--00000000 ?
--1D27ACA0 ?
--E0238B40 ?
F11467D1 (THREADS.NLM|free+58)
--E0238B40 ?
--1D27C010 ?
--1D27AC38 ?
-1DF81C92 (GWCMC.NLM|traceData+DC92)
F1EA6B4F ?
--E0238B40 ?
-1DF81C92 (GWCMC.NLM|traceData+DC92)
--1D27AC44 ?
F1EA6B24 ?
-1DF81C92 (GWCMC.NLM|traceData+DC92)
--1D27AC54 ?
F1EA67C9 ?
-1DF81C92 (GWCMC.NLM|traceData+DC92)
--1D27ACA0 ?
--1D27AC88 ?
F1E96174 (GWCMC.NLM|cmc_realloc+39A9)
-1DF81C92 (GWCMC.NLM|traceData+DC92)
F1E8946D (GWCMC.NLM|cmc_query_configuration+2B5)
--00000000 ?
--00000000 ?
--00000000 ?
--00000000 ?
--00000000 ?
--00000000 ?
--00000000 ?
--1D27C350 ?
--00000004 ?

Additional Information:
The CPU encountered a problem executing code in THREADS.NLM. The
problem may be in that module or in data passed to that module
by a process owned by GWINTER.NLM.

Stranger yet, the user lists in NDS in Groupwise View, in the post office are not there. Users, for the most part are still able to access the internal mail, but internet e-mail and web access remain unusable. We have run both VRepair and DSRepair, and did not get any errors. If anyone has any ideas, please let me know.

Thanks,
Domenick Pellegrini

 

[dpellegrini]

Ok, Here's the resolution. The lock up caused corruption in the GW database.The database had to be recovered and rebuilt using the Groupwise Maintenance, for both the Domain and the Post office. Then rename the WPHost.DB file in the post office. Once this was done, everything seemed to work normally. In our case GW was set to run Client/Server and Direct, in NDS. Our clients still had problems with Notify and 8210 errors. We change the setting in NDS in the PO Object, just to run Client/Server and that resolved the issues with the workstation clients.

The key here was not so much troubleshooting the errors we kept getting, but just rebuilding the database (not a topdown rebuild). Had we been more informed about it, we would have found the solution much faster.

Domenick Pellegrini
dpellegrini@yahoo.com