Scheduler not working after upgrading to SP24

This weekend we upgraded from SP23 to SP24 and everything seemed fine after that. This morning we find that Scheduler is not working at all. It is not submitting any jobs. I have IPL'd the 400 and stopped and restarted the scheduler services. (2Xs) How do I get the scheduler to start submitting jobs again?

Hello,

First a warning: I am not familiar with the 400 as we are a Wintel shop.

However, I have a couple questions:
1. Did the scheduler work this weekend after the upgrade and just stopped sometime later in the weekend or this morning?
2. What does the log for the scheduler kernel say? We use Server Administration Workbench which shows the kernels and their process ID. I can then look at the logs for that process. There should be a log on the server somewhere.
3. Did you review the scheduler kernel definition in the server JDE.INI for errors? Sometimes there are changes required for a Service Pack that need to be made to the current JDE.INI, so it is possible there was an error made during that edit/review process.

Hope these give you some ideas to check.

Jer

Hi Jer,

- After checking, the server stopped working after the upgrade.

- Kernel on the 400 doesn't show any errors.
this is the messages:
Job 432779/ONEWORLD/JDENET_K started on 04/11/11 at 08:30:52 in subsystem
JDEB7334 in B7334SYS. Job entered system on 04/11/11 at 08:30:52.
Scheduler Kernel+(PID = 23 )
Job 432782/QUSER/QSQSRVR used for SQL server mode processing.
Job 432781/QUSER/QSQSRVR used for SQL server mode processing.

- I didn't see in the instructions for service pack to make any changes to the INI file for scheduler. Is this one of those 'read between the line' things that JDE likes to do????

I have placed a service request to Oracle at a P2 level. 3 hours later I have not received a repsonse. WOW is all I can say

I figured the logs would be different. It doesn't look like there are any error messages though.

As for the JDE.INI file changes, "read between the lines" is a good way to put it. There usually isn't anything in the instructions, but I have often found it useful to compare the new generic JDE.INI with my current one to see if there are changes we should be making. Then it is off to research if anything stands out. I looked at our file and I didn't see any changes we made when we jumped from 23ZA1 to 24.1.2.

Can you create a newly scheduled job to see if it will submit? (My favorite is the R0006P). If you set it up to submit in a few minutes that could tell if it is an architecture or it has to do with the existing scheduled records.

Also, is the Job Launcher running (if it exists on the 400)? In our system we have the scheduler kernel running, but there is a process called the Job Launcher which we can pause which will prevent scheduled jobs from submitting without stopping the scheduler kernel itself. In our system, when you start the services (including scheduler kernel) the Job Launcher is automatically started. The job launcher screen shows the next date and time a job will be launched. If the Job Launcher exists on the 400, does it show a good date and time for next submission?


Jer

When you look at the Scheduler Server screen (form exit on the scheduler application) , does it show you the server as started....?


Also check ESU JE9359 , I remember having to apply this ESU when I upgraded to SP24 at a client , will get more specifics in a bit

Ok after looking into this a little more , there is an additional index needed on the F91300 which is what the ESU is delivering. If you look at the debug logs on your scheduler kernel you should see messages related to this...

Only thing is that I only see an ESU for ERP 8.0 and not Xe , your signature says Xe but your logs have JDEB7334 in them , so not sure what your release is.

we are running XE. I was not here when the install and naming went into affect. There was a team of JDE consultants that implemented it.

The server is running. I will look into the ESU.

Thanks for the info


I did receive an email from JDE just asking questions, at 11:30, which I promtly replied back too. They are slow to respond

The Xe ESU will be - JD23936 . I don't have the debug logs from when I encountered the issue , but if you get them from your scheduler kernel I should be able to tell if it was the same issue.

99% this ESU should fix your problem.

Oh yeah - good call. I forgot about that since we added that ESU when we went from 23K to 23ZA1 more than a year ago.

Jer

I downloaded the ESU through change assistant but when I signed in to OneWorld in the JDEPLAN environment to install the ESU, I am getting an error. The error is "Unable to connect to OWBAK_JD23936" I have never seen this error before. Not sure how to install the ESU now

Did you run the JD23936.exe on the deployment server , or was that what you did thru change assistant. Did the folder get created with all the appropriate files.

I would try manually running the exe on the deployment server and then attempt to apply the ESU to a pathcode again logged into JDEPLAN

Yes, I did run .exe and yes it did create the folder for the ESU. When I logged into OW under JDEPLAN, I get the error after I choose the environment and click Next. I have tried PY7333 and PD7333. Neither environment works

Can you try applying to say DV and uncheck the take backup option..?

I tried DV but we don't have that envrionment. So I tried PD and took the check off for backup. Now it works.


Thanks for the help on that one. I am going to build a package and deploy it out the 400. Hopefully this will fix my issue with scheduler.

This ESU has a special instruction to generate the indices on F91300 , so be sure to do that. That is what will fix the issue. The new kernel code looks for an index 11 on the F91300 , the absence of which is what causes it to fail.

the build is taking a very long time for just one object. The client piece went fast, the server piece seems to be stuck and not moving

Hmm check the server build logs..., may be your Enterprise Server is busy with Production processes ?

In all likely hood generating the indices on F91300 should fix your issue. Verify your central object specs to make sure you now have the additional index on F91300. Open UTB , table F98713 (Central Objects datasource) , query on OBNM = F91300 and LNID = 11 , should show 2 records.

Also verify on a FAT client in Table Design by doing a get from PD , if you haven't installed the client package yet.

Stop the scheduler server

Gen the indexes

Start the scheduler server

Hi

I'm pretty sure that if you read the special instructions for the ESU you need to build a full package. This caught me out on an upgrade once.

Thanks
Kieran

Thanks for the ESU. I applied, built and deployed an update package for this. I also did a full save which also does an IPL of the 400 last night. When the system came back up at 1:30am, scheduler started working again.

FYI, Oracle still has not suggested the ESU to me. Plus at 3pm, I esculated the call to a P1 since we use scheduler heavily at night. They did not respond back until 5:30AM (14 hours later). AMAZING, not sure what we are paying for in support but that is just crazy.

I am glad I can count on the people from JDELIST to help me in the right direction

Glad to hear it fixed your problem and that you are back up and running

I'll take that bottle of Glenfiddich now