Netware 5.0 server resets by itself

[spock2054]

The Netware Server, a Compaq Proliant 1600, running Netware 5.00d, keeps disconnecting users with no warning and then resets itself. Sometimes it disconnects users and does not reset itself. It does, however appear that this occurs during disk access as I disabled all logins except myself, Admin, and then tested it. I ran the complete Compaq diagnostics and the hardware passed all the tests. The server is running 2 PIII-600's with the latest BIOS and 768MB of ram. If anyone has any ideas I would appreciate it as this rather an emergency situation.

Thanks

 

[TheLad]

Firstly, apply the latest Service Pack. Then try the following:

SET CLIENT FILE CACHING ENABLED = OFF -----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[spock2054]

I am just a little reticent to "spack" it because the server might go down during the procedure. I entered NWCONFIG to execute it and the server shutdown with multiple abends. I can't even copy the service pack over to the server through windows. It just won't stay up long enough. Any ideas would be welcome! Thanks!

 

[TheLad]

What are the abends received? If they indicate hardware, you may be better reseating your memory and processor(s) for a start. Post the initial abend message from SYS:SYSTEM\ABEND.LOG -----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[spock2054]

Here it is minus the modules loaded info:

Server AT1 halted Wednesday, February 5, 2003 9:31:05 pm
Abend 4 on P00: Server-5.00d-0: Found spinlock to be busy with single processor active

Registers:
CS = 0008 DS = 0010 ES = 0010 FS = 0010 GS = 0010 SS = 0010
EAX = D4BCEB18 EBX = D4BCE000 ECX = D0001004 EDX = 00000058
ESI = D0001038 EDI = D0005700 EBP = D68CCB4C ESP = D68CCB2C
EIP = 0002A283 FLAGS = 00000002
0002A283 83C404 ADD ESP,00000004
EIP in UNKNOWN memory area

Running process: Interrupt service routine (nested count 1)
Interrupted process: netdb 1 Process
Created by: NetWare Application
Thread Owned by NLM: NETDB.NLM
Stack pointer: D68CD5A0
OS Stack limit: D68C9880
CPU 0 (Thread D62E9B80) is in a NO SLEEP state
Scheduling priority: 67371008
Wait state: 3030070 (Yielded CPU)
Stack: -00041AB0 (LOADER.EXE|CurrentYear+352)
--00000046 ?
FC01C258 (SERVER.NLM|FreeMemoryDefault+38)
--D0001004 ?
--00000058 ?
--D19C91C4 ?
--00000058 ?
--D4BCEB18 ?
--D68CCC08 ?
FC01C1AA (SERVER.NLM|Free+122)
--D4BCEB18 ?
--0000000A ?
--D68CCC00 ?
FC054C4C (SERVER.NLM|CheckParameterOrder+4AC)
FC054C36 (SERVER.NLM|CheckParameterOrder+496)
--00000000 ?
--D68CCBE1 ?
--0000000A ?
--00000064 ?
--D68CCC00 ?
FC054C4C (SERVER.NLM|CheckParameterOrder+4AC)
FC054C36 (SERVER.NLM|CheckParameterOrder+496)
--00000000 ?
--D68CCBE2 ?
--0000000A ?
--000003E8 ?
--D68CCC00 ?
FC054C4C (SERVER.NLM|CheckParameterOrder+4AC)
FC054C36 (SERVER.NLM|CheckParameterOrder+496)
--00000000 ?
--00000038 ?
FC054B5C (SERVER.NLM|CheckParameterOrder+3BC)

Additional Information:
The NetWare OS detected a problem with the system while executing a process owned by LOADER.EXE. It may be the source of the problem or there may have been a memory corruption.

 

[TheLad]

Interesting, in your STARTUP.NCF, ensure that the line:

LOAD CPQMPK.PSM

is not there. If it is (should be the first line of STARTUP.NCF), remove it/rem it out. This is the Compaq multi-processor support module and should not be loaded on a single processor machine.

-----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[Lou0686]

Hi,

This is not a good abend to post because it is the 4th abend in a series of who knows how many. You need to post abend # 1. Abend # 4 is no good because the memory has already been corrupted by the first 3.

Lou

 

[spock2054]

This machine actually has 2 PIII-600's. I will post the first abend tomorrow!

 

[spi200]

Hi

My Money is on the memory, so see if you can swap it out. Also unload all NLM and see if it stays up.

regards

David CNE

 

[spock2054]

Here is one from today!

erver halted Friday, February 7, 2003 12:06:39 pm
Abend 1 on P00: Server-5.00d-651: Cache buffer being removed from drive index not found in disk cache.

Registers:
CS = 0008 DS = 0010 ES = 0010 FS = 0010 GS = 0010 SS = 0010
EAX = 2BFF0088 EBX = 0104F4A8 ECX = 00000000 EDX = 00000000
ESI = 00000000 EDI = 2BF8394C EBP = FC5EB4D0 ESP = D05F8D1C
EIP = 00000000 FLAGS = 00000246


Running process: Console Command Process
Created by: NetWare Application
Thread Owned by NLM: SERVER.NLM
Stack pointer: D05F8CB0
OS Stack limit: D05F5040
Scheduling priority: 67371008
Wait state: 5050090 (Wait for interrupt)
Stack: --0000605E ?
--00000006 ?
--9E000004 ?
--00000000 ?
--00000001 ?
D08A0299 (FILESYS.NLM|INWFlushVolume+21)
--00000000 ?
--00000000 ?
D08BDE59 (FILESYS.NLM|DoOldBackouts+505)
--00000000 ?
--9E000004 ?
--00000200 ?
--00000001 ?
--00000000 ?
--D0F25500 ?
--00000001 ?
--00000000 ?
-FC5D92C0 (SERVER.NLM|systemConsoleScreenStructure+0)
--00000000 ?
--D08F3062 (FILESYS.NLM|deadmanCounters+AA92)
--00000001 ?
D08BE26F (FILESYS.NLM|InitializeTTS+3E7)
-FC5D92C0 (SERVER.NLM|systemConsoleScreenStructure+0)
--00000000 ?
--3A36303A ?
--D0E5C0A0 ?
--00000000 ?
-D06EF4CC (CONNMGR.NLM|ConnectionZero+0)
--D0FC18C0 ?
D08A307D (FILESYS.NLM|CloseLock+71)
-D06EF4CC (CONNMGR.NLM|ConnectionZero+0)
--00000000 ?

Additional Information:
The NetWare OS detected a problem with the system while executing a process owned by SERVER.NLM. It may be the source of the problem or there may have been a memory corruption.

 

[Lou0686]

Hello again Spock2054,

You cut off the top part of the abend, which is important. Was it an NMI or Page Fault abend? But considering that it is cache related, it sounds like bad memory to me too.

Lou

 

[spock2054]

I just checked it and the only thing missing from the beginning of the abend is the "S" in server.