DUAS 5: many SAP jobs abort at the same time

Document ID : KB000087152
Last Modified Date : 14/04/2018
Show Technical Document Details
Issue:
Affects Release version(s): 5

Error Message :
Universe.log: 
------------------- 
<< 2012-11-05 11:12:30 4391858/uxcdjsrv /u_cdj_envoi_buffer /000000000 - (index = 32766)/send error u_rep_req: Errno syserror 32: Broken pipe 
<< 2012-11-05 11:12:41 15860182/ /UXSAP_TermProcess /000000000 - Receive signal 15: stop all jobs for numproc=1999618 codug=S_PRD_VAN 
<< 2012-11-05 11:12:44 13631688/ /UXSAP_TermProcess /000000000 - Receive signal 15: stop all jobs for numproc=1999952 codug=S_PRD_VAN 
<< 2012-11-05 11:15:22 4391858/uxcdjsrv /u_cdj_envoi_buffer /000000000 - (index = 32766)/send error u_rep_req: Errno syserror 32: Broken pipe 
<< 2012-11-05 11:18:16 4391858/uxcdjsrv /u_cdj_envoi_buffer /000000000 - (index = 32766)/send error u_rep_req: Errno syserror 32: Broken pipe 
<< 2012-11-05 11:19:58 13631688/ /u_api_call_agtsap_gu/000000000 - Network error receiving data from Dollar Universe Manager for SAP: Errno syserror 4: Interrupted system call (recv returns error) 
<< 2012-11-05 11:19:58 13631688/ /u_api_jobsap_create /000000000 - fails in communication with SAP agent [-1] 


sapjcs.log: 
--------------- 
2012-05-11 06:45:08 # u_build_joblog # Unable to get spool from step 1 
2012-05-11 08:52:01 # agt_jnl_del_job # Job (VAN_CUSTOMER_ASNUPDATE_US ) (08512500) not found in jnl file (0) 
2012-05-11 09:31:49 # uxsap_api_fla # Task not found (error code -404) 
2012-05-11 09:31:49 # uxsap_trt_fla_error # error [object not found] for launch [JOBNAME:VAN_BILLING_DUE_LIST_IT_1031 JOBCOUNT 09305700 USER:XSLANKIP STATUS:P] 
2012-05-11 09:36:12 # uxsap_api_fla # Task not found (error code -404) 
2012-05-11 09:36:12 # uxsap_trt_fla_error # error [object not found] for launch [JOBNAME:VAN_BILLING_DUE_LIST_IT_1031 JOBCOUNT 09360300 USER:XSLANKIP STATUS:P] 
2012-05-11 09:45:20 # u_build_joblog # Unable to get spool from step 1 
2012-05-11 10:07:24 # agt_jnl_del_job # Job (VAN_CUST_ASN_CANADA ) (10054100) not found in jnl file (0) 
2012-05-11 10:12:25 # uxsap_api_fla # Task not found (error code -404) 
2012-05-11 10:12:25 # uxsap_trt_fla_error # error [object not found] for launch [JOBNAME:VAN_BILLING_DUE_LIST_IT_CDR_1031 JOBCOUNT 10120500 USER:XSLANKIP STATUS:P] 
2012-05-11 10:45:15 # u_build_joblog # Unable to get spool from step 1 
2012-05-11 11:31:42 # agt_process_client_reques # Network error: Errno syserror 32: Broken pipe (send returns error) 
2012-05-11 11:31:43 # agt_process_client_reques # Network error: Errno syserror 32: Broken pipe (send returns error)

Patch level detected:Manager for SAP 4.4.1
Product Version: Dollar.Universe 5.6.0 FX25010 + Manager for SAP 4.4.1

Suddenly many SAP jobs abort at around the same time.

Jobs stop running on Dollar Universe, they get stuck in launch wait status.
Suddenly they start running again, and then after 2-3 minutes they get stuck in launch wait again.
After restarting the dollar universe node, jobs still do not run.

The SAP manager is recycled but it does not resolve the issue.
Environment:
OS: IBM AIX
Cause:
Cause type:
Defect
Root Cause: It is likely to be due to the fact that the SAP manager (uxagtsap) reaches a memory limit due to a memory leak in case of a lot of activity.
Resolution:
1. If it's not already the case, we recommend to upgrade to SAP manager patch FX24937A, which fixes a memory leak issue.

2. If the OS is AIX, increase the maximum memory allowed for the SAP manager. This will allow the SAP manager to use up to 1G memory.

 
LDR_CNTRL=MAXDATA=0x40000000
export LDR_CNTRL


Fix Status: Released

Fix Version(s):
Component: Manager.For.SAP 4.4.1
SAP manager patch FX24937A - Available
Additional Information:
Workaround :
Rename the uxsapjnl.dat data file, and restart the SAP manager.