jdequeue.exe process dying

[bknowles]

I'm having a problem with a Windows 2003 Standard logic server losing a single jdequeue.exe process. Nothing in the
Windows event logs.

I'm seeing the following in 2 of the jde_xxx.log files:

File #1
452/344 Thu Oct 07 10:19:49 2004 IPCPUB.C3135
process 452 <jdequeue.exe> registered in entry 46

452/344 Thu Oct 07 10:37:35 2004 IPCWNT.C1580
IPC3100014 - Mutex <index=59> was abandoned by it's previous owner. Taking over lock.

File #2
2872/3376 Thu Oct 07 10:19:49 2004 IPCPUB.C3135
process 2872 <jdequeue.exe> registered in entry 43

2872/3376 Thu Oct 07 10:37:35 2004 jdbodbc.C5557
ODB0000207 - SQLExecute failed

2872/3376 Thu Oct 07 10:37:35 2004 jdbodbc.C5557
[Microsoft][ODBC SQL Server Driver][SQL Server]Transaction (Process ID 144) was deadlocked on lock resources with another process and has been chosen as the deadlock victim. Rerun the transaction. - SQLSTATE: 40001

2872/3376 Thu Oct 07 10:37:35 2004 JDB_DRVM.C938
JDB9900401 - Failed to execute db request

2872/3376 Thu Oct 07 10:37:35 2004 JDEQUEUE.C804
JDB_SelectKeyed F986110 failed with rcode = 0

2872/3376 Thu Oct 07 10:37:35 2004 IPCPUB.C3515
API ipcSawUnregisterProcV1 : process 2872 unregistered in entry 43



Does anyone have a clue what could be causing this?

Thanks,
Barry Knowles

 

[Alex_Pastuhov]

This particular one was killed because of deadlock.

Such deadlocks are ultimately caused by design bugs in OW or Applications, but you can minimize the frequency of this happening by improving the SQL performance and thus making your DB transactions as short as possible.

 

[bknowles]

I figured out what was causing the problem. In B7332, F986110 had a clustered index. In ERP8 it does not. Even
re-generating the table through OMW did not create a
clustered index. I manually clustered it and the problem is
gone. Pulling up submitted jobs is also much faster.