XI3.0 Scheduled Reports Going To Pending

scheduling XI Scheduler
1

I have a brand new install of BOXI 3.0 with FP1 and SAP Integration KIT.

When I schedule a report in WebI (using eFashion as universe) as Run Now or Run Once the report goes in as pending and stay as pending.

I have tried restarting the SIA, recreating the WebI processing server and recreating the Adaptive job server to no avail.

I’ve seen blogs suggesting issues with DB Connectivity associated with this symptom however I’m discounting that on the grounds that the SQL Server instance for the Repository DB Server is on the same box as the Business Objects install

Another item that may be relevant is that this machine is quite slow, with only 1GB of RAM, could that be resulting in some sort of time out issue?

In an attempt to resolve I paused Instance manager, cleaned off the old jobs and added the following switches to the Job Servers:

-maxChildReqs 1 -childProcTimeOut 30 -checkChildForNonExistence 1

I then stopped the SIA and started it and confirmed from the logs that the job server log was clear.

I then scheduled one report to run ‘Now’. It is persisting in a state of pending.

The logs show:

trace message: Parent_impl::wait() 13000
trace message: Parent_impl::findChild: 1884
trace message: Parent_impl::findChild: could not find 1884
trace message: SI_IInfoProcessing_implMixedProc::childRequest() waiting for child: 1884
trace message: Parent_impl::wait() 13000

This then repeats for a total of 5 times then does:

trace message: Destinations::loadProperty(): ISPROP_IPROC_CURJOBS.
trace message: jobserver::countAndCleanupJobsHelper(): checkChildForNonExistence[1], forceGetLoadtrue, timeElapsed60020, currentLoadTotal[0]
trace message: jobserver::countAndCleanupJobsHelper(): returning currentLoadTotal[0]
trace message: Destinations::loadProperty(): ISPROP_IPROC_MAXJOBS.
trace message: CServiceAdmin_impl_base::getNVMetrics
trace message: Parent_impl::findChild: 1884
trace message: Parent_impl::findChild: could not find 1884
trace message: jobserver::newIJob(): [Couldn’t get IJob interface or writing IAudit: Pipe exception. Reason: jobserverchild (WebIJavaSchedulingService, 0, Timeout waiting for Child 1884 to register (120seconds).

Cycles through the parent_impl again and then starts again for the same job with a new child number.

It looks like it’s processing through but is instead stuck in and endless loop.

Associated Error in Event Log:

“Subprocess (Job Server Child) could not be started. Reason: Couldn’t get IJob interface or writing IAudit: Pipe exception. Reason: jobserverchild (ReplicationSchedulingService WebIJavaSchedulingService, 0, Timeout waiting for Child [4116] to register ([120]seconds).”

Does anyone have any ideas?

Damocles :uk: (BOB member since 2006-10-05)

2

Hi,
Are you using a specific port for the Jobservers?

thanks
Suresh

suresh12 (BOB member since 2009-03-04)

3

Try changing the Command Line Parameters to
-loggingPath “D:/BOXI/BusinessObjects Enterprise 12.0/Logging/” -jsTypeDescription “Desktop Intelligence Job Server” -objectType CrystalEnterprise.FullClient -maxDesktops 0 -type OUTPROC

Aparentley the jobeservers assign tasks to themselves and can mean that jobs sit at pending for a while. OUTPROC at the end of the command line stops this. I’m not sure the benefit of the default setting

Xavier :ireland: (BOB member since 2006-12-18)

4

Hi

I have a similar problem - but have XIR3.1 - my schedules are stuck in pending and as a result the following scheduled report are failing.

The only way to resolve it is to restart the SIA. We have upped the number of concurrent jobs (at first we thought it was the number of jobs going out) - but today we dont have many at all going out.

If you look at the error on the following scheduled failure it is -

An internal error occured while calling ‘openDocumentMDP’ API. (Error: ERR_WIS_30270)

Thanks

Karen

Babelfish :uk: (BOB member since 2006-07-28)

5

This seems to be a generic error.
have a look at this thread, it may throw some light on the issue.

Xavier :ireland: (BOB member since 2006-12-18)

6

Damocles, have you made any progress on this behavior? We are seeing the same exact behavior and have exhausted all avenues, except logging a case.

kmspsu93 :us: (BOB member since 2006-04-06)

7

Hi,

I’ve got the samee problem with XI 3.1, using webi scheduler: every report stay in status pending.

Below SAP solution found today on its portal;
I applied this solution and now everything works fine !!
Hope this helps:

Symptom

Scheduled WebIntelligence reports remain stuck in ‘pending’ status after installing Microsoft windows patches of April 2009. Restarting the Adaptive Job Server, Central Management Server or Server Intelligence Agent has no effect.
The Event Viewer logs show the following entry: Unable to start the sub-process (Job Server Child). Cause : Couldn’t get IJob interface or writing IAudit: Pipe exception. Reason: jobserverchild (WebIJavaSchedulingService ReplicationSchedulingService, 0, Timeout waiting for Child [4716] to register ([120]seconds).

Reproducing the Issue

Install Business Objects Enterprise XI 3.x on a French Windows 2003 SP2 32bits server
Install the following Microsoft windows update patches of April 2009:
KB923561, KB925336, KB952004, KB956572, KB959426, KB960225, KB960803, KB961373, KB967715
Schedule a WebIntelligence report
The report remains in Pending status
Business Objects Enterprise XI 3.1
Windows 2003 SP2 32bit (French)
Microsoft windows patches of April 2009

Cause
With Microsoft windows patches of April 2009, the maximum Heap Size for java processes has been reduced to under 1000MB.

Resolution

Apply the following steps to fix the issue:

Launch the Central Management Console
Select Servers
Right click on <server_name>.Adaptive Job Server and select Properties
Add the switch -javaArgs “Xmx900m,Xincgc,server” in the command line of the server
Click on Save and Close
Restart the <server_name>.Adaptive Job Server

sarce :it: (BOB member since 2008-04-28)