TAM locks are killing me

Here's what we've got...

OW (Base, SP 19.1) World A7.3 X3 - Coexistant
AS/400 ES (720...soon to be an 820) running V5R1
Deployment Server (WinNT, MS SQL) - Central Objects
2 Citrix boxes (W2K, MetaFrame, WTS)

Here's our problem...

I have the following log from a job that went ERROR status on the AS/400 as soon as it was submitted...

435 Wed Oct 23 10:17:07 2002 runbatch.c382
INITIALIZING RUNBATCH job: 84092

435 Wed Oct 23 10:17:07 2002 ipcpub.c3121
process 435 <B7333SYS/PRINTUBE> registered in entry 7

435 Wed Oct 23 10:17:07 2002 runbatch.c614
Generating Joblog for UBE, refer to joblog of job 337423/JDE /R0006P for additional information.

435 Wed Oct 23 10:17:07 2002 runbatch.c632
Startup for User=ADBRES0, Env=PD7333, Job#=84092

435 Wed Oct 23 10:17:18 2002 ipccrt.c178
IPC2100002 - aPD7333aspecfileadstmpl.ddb exists, but type is not eIPCSWMRLock.

435 Wed Oct 23 10:17:18 2002 jdetama.c407
TAM0000097 - Unable to create TAM lock for table aPD7333aspecfileadstmpl.ddb.

435 Wed Oct 23 10:17:18 2002 jdetama.c600
TAM0000103 - Unable create lock for table /PD7333/specfile/dstmpl.ddb for WRITE.

435 Wed Oct 23 10:17:18 2002 runbatch.c2087
TAMInitOrCreate failed for eSpecType 10 in InstallSpecsForUBE

435 Wed Oct 23 10:17:18 2002 runbatch.c937
SpecInstall failed in main

435 Wed Oct 23 10:17:18 2002 ipcpub.c3432
API ipcSawUnregisterProcV1 : process 435 unregistered in entry 7

435 Wed Oct 23 10:17:18 2002 runbatch.c1379
PrintUBE failed, refer to joblog of job 337423/JDE /R0006P for additional information.

435 Wed Oct 23 10:17:18 2002 runbatch.c1395
Job of type UBE failed with rcode = 0

The only way I have found to resolve this issue is to stop and restart services on the AS/400 (I've also rebooted the Deployment server too). I would like to find out why this is happening. Why are we getting these locks? I'm not sure what other information I need to offer up in order for someone to help me out here, so ask away, and I will do my best to let you know what's going on.

This problem used to happen at least once a week. I originally attributed this to a possible V4R4 to V4R5 upgrade that may have missed a CUME update, as this problem went away when we upgraded from V4R5 to V5R1. We've been on V5R1 since 9/25/2002 and have not had this problem occur until today. Jobs were running fine this morning until about 10:00 a.m. and the all of the sudden began ending in ERROR status right away after being submitted. Nothing was being installed, changed, modified, deleted, etc from the AS/400 or Deployment server during this time.

About the only thing I can think of is a problem we've been having with our network backup application (Backup Exec), which does back up our deployment server, but it is only supposed to backup a copy of the production database. We've had problems with Backup Exec and our Lotus Notes (WinNT) with database locks, and was wondering if anyone else is experiencing problems with locks on the Deployment server.

I would greatly appreciate any help anyone can offer.

Thanks,

Steve,

I have seen something like this, or at least very close to it. It occured
everytime that a UBE was submitted during the SQL Server backup of the
Central Objects. To resolve this issue, we had to reboot the SQL Server
after the backup was completed. Also, we had to make sure that no jobs were
submitted during the SQL db backup.

Hope this helps,


Andy




Consultant with clients
at B7331 - Xe
Various Deployment Servers, Citrix and JAS

Call JD Edwards. This is a bug in SP19.1 and I had a call open for 2 months
on it. They now have a fix for it. I'm not sure if it's available in a SAR
yet, but 2 weeks ago they sent us an update to the jdeipc service program,
and it worked.



Xe, SP 19.1_E1, Update 6, ES=AS/400 V5R1, CO=AS/400, Thick & Citrix Clients

Thanks Andy...

The only problem is that there is a backup that runs at 10p.m. at night in SQL Server, and it completed in about 15 minutes. This morning, jobs were running fine until about 10:15-10:30 this morning. Nothing was going on in SQL Server, other than normal JDE OW things. No backups or maintenance stuff. I haven't looked, but I will, but I don't think there are any scheduled jobs that run nightly, at 10p.m., in OW, but if there are, and they would cause this locking problem, why would things work fine for awhile and then just stop working all of the sudden.

Since you have an AS/400, you might think about moving the central objects to the AS/400. Then you don't need SQL server (or license fee) on the deployment server.

Another thing that locks up UBE's is deploying an update package.

What does the joblog for "Generating Joblog for UBE, refer to joblog of job 337423/JDE /R0006P" show?

Have you tried setting on the debug for the server?

Notes from our issues database...we had exact same problem...


This issue was resolves after bouncing Ent Server, however, see the bottom for a issue I found on KG. To prevent this from happening in the future, the following changes have been put in effect:

Under JDEIPC section in the jde.ini, the following lines were added:

maxNumberOfSemaphores=500

& changed the IPCStart value to 7000

Previously, in the earlier releases when launching UBEs to servers, the TAM Specifications are locked with a read lock for the duration of the UBE. This is to prevent a very small window of opportunity where more than one process can be
writing to the same file at the same time. Consequently, any UBEs submitted with a spec install will wait until the running UBEs are complete. But now new codes (since B733) has been added to provide SWMR locks on the following tam
spec files when running on the servers: GBRSPEC, GBRLINK, RDASPEC, RDATEXT, DSTMPL. UBE's are no longer locked the entire time they are running, they only
lock the files as they are used. So if you are wondering why certain specs are locked as in your case, the above provide the explanation.



35785 Thu Jun 6 12:54:49 2002 dbdrvsql.c251
OS40002001 - Error Type: OTHER enum: 9 - Could not bind to the QZDMMDTA SRVPGM or could not find one of its exports!

ered in entry 12

35785 Thu Jun 6 12:54:48 2002 runbatch.c632
Startup for User=XXXXXX, Env=PY7333, Job#=23394

35785 Thu Jun 6 12:54:49 2002 ipccrt.c178
IPC2100002 - aPY7333aspecfileagbrlink.ddb exists, but type is not eIPCSWMRLock.

35785 Thu Jun 6 12:54:49 2002 jdetama.c407
TAM0000097 - Unable to create TAM lock for table aPY7333aspecfileagbrlink.ddb.

35785 Thu Jun 6 12:54:49 2002 jdetama.c600
TAM0000103 - Unable create lock for table /PY7333/specfile/gbrlink.ddb for READ.

35785 Thu Jun 6 12:54:49 2002 ipccrt.c178
IPC2100002 - aPY7333aspecfileagbrlink.ddb exists, but type is not eIPCSWMRLock.

35785 Thu Jun 6 12:54:49 2002 jdetama.c407
TAM0000097 - Unable to create TAM lock for table aPY7333aspecfileagbrlink.ddb.

35785 Thu Jun 6 12:54:49 2002 jdetama.c600
TAM0000103 - Unable create lock for table /PY7333/specfile/gbrlink.ddb for READ.

35785 Thu Jun 6 12:54:49 2002 ipccrt.c178
IPC2100002 - aPY7333aspecfileagbrlink.ddb exists, but type is not eIPCSWMRLock.

35785 Thu Jun 6 12:54:49 2002 jdetama.c407
TAM0000097 - Unable to create TAM lock for table aPY7333aspecfileagbrlink.ddb.

Dave Wilson

OW Xe SP19 Upgrade 3
AS/400 V5R1
NT / Citrix