Service issue

Hi

Can anyone suggest how to resolve the below issue:

jsh ~ -->tSA 2
tSA 2
<como>
<agent>2</agent>
<processid>51249188</processid>
<portno>9</portno>
Agent 2 started 15 DEC 16 07-44-31
Agent's Process id 51249188
<servername>EWRQWER</servername>
Running on server EWRQWER PortNumber 9
Next Service Current Service
<service name = BNK/XXX>
<process name = BNK/XXX>
</process>
No Service Allocated for this Agent
</service>
Agent stopped
</como>
___2_15 DEC 2016_07:44:31:783_SELECT F.RECORD.LOCK WITH K.USER EQ "" Selected=0 time=0secs
jsh ~ -->

Steps taken to overcome this issue

1. related F.BATCH.STATUS and F.LOCKING and JOB.LIST records were cleared
2. T24.SESSION records were all cleared.
3. TSA.STATUS was cleared and TSM restarted.

But still same issue is thrown.

Hi
As far as I see, you're starting tSM in DEBUG mode.
What's in the output of tSM session?

Hi VK

Thanks for the response. Issue got fixed. this is really crazy but no idea why.

Allow me to shed some light on how the single thread routine works:
1. Single thread service checks for a file in a folder. If file found, then starts a multi-thread service to process the file.
2. if the multi-thread service is not found in environment, then single thread service passes parameters using F.WRITE to create batch and service and starts the service. (I ignored using OFS becuase of below issue - again no idea why that happened - People I requested for help were also confounded)***
3. Issue was thrown with "No Service Allocated for this Agent" message while starting the new multi-thread service.
4. this issue did not occur for other services that were running parallely - but only for the service i created using single thread.

Now the current issue got fixed when i also passed the PROCESS.STATUS and JOB.STATUS values as 0 in F.WRITE while creating the batch in singlethread.

OFS issue - which occurred while running as service
******

Debug screen.

0085 Y.OFS.MSG.START.SER = "TSA.SERVICE,XXX/I/PROCESS,INPUTT/null,BNK^XXX.":I:",SERVICE.CONTROL::=START"
jBASE debugger->S
0086 Y.OFS.RES.START.SER = ''
jBASE debugger->V Y.OFS.MSG.START.SER
Y.OFS.MSG.START.SER : TSA.SERVICE,XXX/I/PROCESS,INPUTT/null,BNK^XXX,SERVICE.CONTROL::=START
jBASE debugger->S
0087 Y.OFS.COMMIT.START.SER = ''
jBASE debugger->S
0088 CALL OFS.CALL.BULK.MANAGER('GENERIC.OFS.PROCESS', Y.OFS.MSG.START.SER, Y.OFS.RES.START.SER, Y.OFS.COMMIT.START.SER )
jBASE debugger->S
0090 RETURN
jBASE debugger->V Y.OFS.MSG.START.SER
Y.OFS.MSG.START.SER : TSA.SERVICE,XXX/I/PROCESS,INPUTT/null,BNK^XXX,SERVICE.CONTROL::=START
jBASE debugger->V Y.OFS.RES.START.SER
Y.OFS.RES.START.SER : Sorry, you don't have permissions to access this data
jBASE debugger->

Check the message due to which record does not get updated - Sorry, you dont have permissions to access this data.

But the same coding is executed successfully in a program in same environment.

Hi
From a cursory glance it looks that it might be better to use only multi-threaded service (where the check of file presence will be done in ".SELECT" routine)...

Hi VK

We are doing two selections here.
1. One is performing a select on the folder for the number of records. there can be any number of records.
2. Selecting the content of the record (with OFS string) to be passed on to the BATCH.JOB.CONTROL.

And since .SELECT is triggered only once in a multithread, then we will have to use the first select in the .LOAD routine while the service is set to AUTO. Now we will have to two selected values kept in loop waiting for the previous select to complete. This would be sequential but we want parallel processing.

But if we have to use the two selects mentioned above then we should have a developed system in T24 to rebuild job list dynamically, which is not possible.

Can anyone suggest how to resolve the below issue:

jsh ~ -->tSA 2
tSA 2
<como>
<agent>2</agent>
<processid>51249188</processid>
<portno>9</portno>
Agent 2 started 15 DEC 16 07-44-31
Agent's Process id 51249188
<servername>EWRQWER</servername>
Running on server EWRQWER PortNumber 9
Next Service Current Service
<service name = BNK/XXX>
<process name = BNK/XXX>
</process>
No Service Allocated for this Agent
</service>
Agent stopped
</como>
___2_15 DEC 2016_07:44:31:783_SELECT F.RECORD.LOCK WITH K.USER EQ "" Selected=0 time=0secs
jsh ~ -->

Hi
as far as I see, you have executed in another session START.TSM -DEBUG. What is there in that session? Something like "manually launch tSA nn"... Is tSA 2 on that screen?

Hi
what exactly is the issue? Can't read murky text with my head turned by 90 degrees  :)))

Line 48 , Source FGB.CI.ORD.CLEAR.POST - what's there? Looks like you're trying to deal with non-valid file descriptor.
 

this is the code which starts from 48 line.
ERR.LIST<-1> = ERR.DTLS
WRITESEQ ERR.LIST APPEND TO ORD.LOG.FILE ELSE NULL
WRITESEQ STR('-',80) APPEND TO ORD.LOG.FILE ELSE NULL
WRITESEQ CHARX(12) APPEND TO ORD.LOG.FILE ELSE NULL
CLOSESEQ ORD.LOG.FILE

Hi
do you have in your code the statement looking like OPENSEQ ... TO ORD.LOG.FILE ?

Hi VK,
The mentioned statement looking like OPENSEQ ... TO ORD.LOG.FILE is not present in this routine but in LOAD routine below is present.
kindly check


* Opening all the necessary files for the subroutine to continue
*
FN.ORD = 'F.OFS.REQUEST.DETAIL' ; FV.ORD = ''
CALL OPF(FN.ORD,FV.ORD)
*
ORDLOG.FP = 'ORD.LOG'
ORDLOG.NAME = 'ORD.DELETED.DETAILS'
OPENSEQ ORDLOG.FP,ORDLOG.NAME TO ORD.LOG.FILE ELSE
CREATE ORD.LOG.FILE ELSE ABORT ;* Create the File if not present
WEOFSEQ ORD.LOG.FILE
END

RETURN

Hi
Do you have ORD.LOG.FILE defined in your COMMON area?

Actually, ORD.LOG.FILE better to be opened not in LOAD but in "record" routine - now it is truncated by every agent in "parallel"... ; you need to add something like "-TNO" to its name to avoid 2 agents to simultaneously write to the same file.


 

Hi,
ORD.LOG.FILE is defined in common area.
Actually, this is defect issue and I am doing it for first time.
It will be helpful if you suggest how to write it using "-TNO".

This will add to file name the session number that is unique for all currently working sessions.

Hi VK,
thanks for the help.

Hi
A core t24 error; subject for Temenos support...

What looks strange to me:
- DATE.RESOLVED........ 00 APR 2023  for previous error;
- TIME.DATE............ (empty and 00 APR 2023)
- CONTRACT MM222775 (too short)

Hi VK,

thanks for your help
How to process this next?
Can we get source code of it

Hi
ask Temenos support... but they don't supply sources (except one-time fixes)

Hi VK,

Can you pls help me with one issue in that under ECB application one nostro account with HVT.FLAG as "yes" is used to book LD,FT, MM
The issue is in this account arround 1900 Un auth FT present but when searching these FTs they are missing

Due to which i cant able to delete them?
Can you suggest what has to be done next?

Thanks for your continous support