Views:

Summary

The scheduler's failure to connect to Syncese on a busy master server prevents jobs from starting and cites return code 8 during the set job status phase.

 

 

Symptoms

 

 

If the scheduler module encounters a send/receive failure while attempting to connect to a master server's local Syncese process, the job log will display the following error:

 

 

SNBSCH1101E CM API function:CM_AP_CONNECT_PLIM return code: 12032 on socket: 6

SNBSCH5693J connecting to syncese at <Master IP> failed (rc = 12032)

SNBSCH5705J Set job status: job <Job Name> (sched_id ########) failed (rc = 8)

The corresponding *.scl (Scheduler) log will display the following error message:

Tue Aug 12 23:00:00 2008 SNBSCH5692I Connecting to syncese at <Master IP>........

Tue Aug 12 23:02:00 2008 SNBSCH1101E CM API function: CM_AP_CONNECT_PLIM

return code: 12032 on socket: 6 Tue Aug 12 23:02:00 2008 SNBSCH5693J connecting to

syncese at <Master IP> failed (rc = 12032)

In the above message, 12032 represents a send/receive timeout of two minutes.

The following excerpt from the *.esl (Syncese) log will reveal that a new Syncese process was launched but became busy trying to connect to the logger module and could not respond to the scheduler's request:

Tue Aug 12 23:00:00 2008 SNBGEN1110I LogFile (<SSPRODIR>/logs/000044e4.esl) opened by Module(syncese) Tue Aug 12 23:00:00 2008 SNBGEN1111I Module Build ID:[arch(sun56) ver(3.0) build_time(01:13:24) build_date(Jul 11 2008)]

Tue Aug 12 23:00:00 2008 SNBGEN1113I Process Info:[pid(17636) user(root) cmdline(syncese)] Tue Aug 12 23:00:00 2008 SNBGEN1112I Env:SSPRODIR(<SSPRODIR>)

Tue Aug 12 23:00:00 2008 SNBGEN1112I Env:SSICMAPI(-sbs 65535 -a 6123 -tnd -hn <SSPRODIR>)

Tue Aug 12 23:00:00 2008 SNBGEN3066D ms_register_module_str: listening on TCP port 64268

Tue Aug 12 23:03:47 2008 SNBESE5709I Logger initialization failed (rc = 10060)

The connection to the logger process fails after almost four minutes and returns a 10060 connection timeout message. Typically, logger connection issues are not fatal, but in this case the logger initialization failure prevented Syncese from responding to the scheduler requests within the scheduler's timeout threshold.

The logger.* (Logger) logs show no evidence of the connection attempt, since it had reached the default connection limit of 256.

 

Resolution

 

 

To resolve this issue, increase the logger's connection limit from 256 to 512 and restart BEX on the Master Server. The logger connection limit is set by modifying the -C <value> (upper case 'C') for SSLOGGER on the Master Server.

 

 

In DPX version 3.0 and later, this change can be made through the Configure Enterprise window on the DPX management console by performing the following procedure:

  1. Right-click on the Master Server Node and select Manage Configuration.
  2. Select logger (Central Logging Dispatcher) in the right pane's module list.
  3. Select Advanced.
  4. Enter 512 in the box next to the -C.
  5. Ensure that the checkbox next to the -C is selected.