• JDELIST is celebrating its 20 year anniversary today!
  • Introducing Dark Mode! Switch by clicking on the lightbulb icon next to Search or by clicking on Default style at the bottom left of the page!

Killing Runbatch on NT 4.0

cdawes

VIP Member
The inability of OneWorld to consistently kill the Runbatch.exe process
occassionally causes OneWorld to be down following a cold backup.

Currently we do cold Oracle backup's that stop the JDE Queue services, waits
2 minutes for all the kill queue's, stops JDE Network services and then
stops Oracle so the database export can start. After the export the backup
script restarts Oracle, waits 2 minutes for Oracle to come up, starts JDE
Network services waits 2 minutes and then starts JDE Queue services. All
this works great most of the time. However, if a UBE is running at the time
the JDE services are stopped then Runbatch.exe will still be running and
will not be killed when the JDE services are stopped. If runbatch.exe is not
killed then the JDE services do not start the correct number of processes
(the services do start) and OneWorld is effectively unavailable in the
morning..

Has anyone else experienced this? Does anyone have an easy solution for this
(ie. an easy utility to automatically kill an NT process at a predetermined
time). We're looking at using the MKS toolkit to try and deal with this but
hopefully a OneWorld solution can be found (I do have killimmediate=1 in the
server ini but this doesn't solve the problem). I have found two SAR's that
relate to this issue (3508897 and 4089313) but they obly state that the
problem has been corrected in B9.........not too much use to me right now.

Thanks

Colin

Colin Dawes, MSc
City of Guelph
B7332 SP15.1_009, Oracle 8.1.6.3.4, NT 4.0, Fat, WTS, Citrix

P.S. Happy Canada Day!!!



Colin Dawes, MSc
City of Guelph
B7332 SP13.1, Oracle 8.1.6, NT 4.0, Fat & WTS
 

cdawes

VIP Member
Sorry, forgot one important thing. I wanted to be able to report on the
outstanding processes and confirm the successful shutdown of OneWorld and
all the associated processes at the NT level, not through the JDE logs. I
know 'at /every [ ] time [ ] kill runbatch.exe' will do the job but can I
confirm this by querying the NT processes and reporting on them?

Colin




Colin Dawes, MSc
City of Guelph
B7332 SP13.1, Oracle 8.1.6, NT 4.0, Fat & WTS
 
G

Guest

Guest
Colin,

We do COLD backups here to on W2K box. As we're in implementation I not
experiencing any UBEs running when backups run.
I'm slightly confused by your reference to "export" during COLD backup as
database needs to be up for Export utility to run.
Having said all this, have U looked at tools available in NT Resource Kit
which might help resolve your problem. PULIST [RK] utility lists the
process ids of running processes and KILL utility can be used to stop
processes. Whether this is "acceptable" from JDE's perspective is another
question but some more experienced folk on this list may enlighten us!

A useful reference book which may interest you is "Windows NT Shell
Scripting" by Tim Hill, ISBN 1-57870-047-7

HTH,
Sean O' Neill - [Digest Mode]
Organon (Ireland) Ltd.
B7333 - SP15.1 - NT W2K SP1 (NTFS) - Oracle 8.1.7
SANDBOX: None to date ;)
 

pareano

Member
After our cold backup we have scripts to auto-reboot the box. This solves the problem of hanging killques and runubes.

Paul

Paul Areano
One World B7.3.3.2 SP 15_007 Intel NT Oracle
CNC/Applications/Etc. etc. etc.
 

cdawes

VIP Member
Thanks Sean,

PULIST.EXE, KILL.EXE and TLIST.EXE from the resource kit will do the job
nicely. We're only doing this temporarily as we have decided to install the
MKS toolkit (allow Unix like monitoring and commands on Windows NT) on the
Enterprise server.

Colin





Colin Dawes, MSc
City of Guelph
B7332 SP13.1, Oracle 8.1.6, NT 4.0, Fat & WTS
 
Top