JobServerChild.exe rises to 100% CPU usage (boXI r2)

[Former Member]

On a single server install of business objects XI release2 (deski job server 11.5.9.1076) we are seeing JobServerChild.exe processes rising to 100% CPU usage and remaining there unless manually terminated. Typically we end up with 4 processes (on a 4CPU machine) each trying to use a whole CPU. I've kept track of some of these over a few days and the same process is remaining alive and increasing the resources used until the whole machine is unusable (if left alone).

I understand that these processes are related to scheduled jobs and have checked these, but have not been able to link it to any particular job. Most of our scheduled jobs are running overnight (3-4 in working hours) and running reports produced in deski to be saved on the server or mailed to users. I've spread this schedule out so that no more than 2 reports run in any 15 minute period but this hasn't impacted the problem noticably. All of these reports will refresh manually in 2 minutes or less.

Is there something that can be configured to prevent this occuring?

Apologies if I've missed anything obvious, google keeps returning links to old, inaccesible forum pages.

Many thanks for any input.

[Former Member · Accepted solution]

You may want to change the MaxChild Reqs to 1 by entering the following command line for the CR Job Server

-maxChildReqs 1

This will ensure that the OS terminates unnecessary JobServerChild executables that remain in memory for simultaneaus requests...reducing the memory and processor utilization.

Below are additional parameters configurable for the CR Job Server

:

u2022 -maxJobs

Sets the maximum number of concurrent jobs processed by one job

server.

u2022 -type

Presets the job server type. Can be OUTPROC or MIXEDPROC.

OUTPROC causes the job server children to shutdown after every

job (like CE10 report job servers), MIXEDPROC causes the job

server children to shutdown after every 100 jobs.

u2022 -maxChildReqs

Sets the maximum number of jobs processed by one job server child

before it shuts down and restarts (in case you want a to use a

number between 1 [OUTPROC] and 100 [MIXEDPROC])

Job server children will also automatically shut down if left unused for a period longer than 2 hours.

Antonio

[Former Member]

Wondering if this has come up for anyone again? We have situations where the server stays pegged at 100% CPU with multiple JobServerChild.exe processes eating up around 13% CPU each.

Moving the reports to webi is not an option for us because the reports are written out to a filesystem for consumption by a custom dashboard. Webi reports do not have the option to write scheduled reports to a filesystem.

[Former Member]

I've found the same problem and suspect it is responsbile for causing slow response times in our BOXIr2 CMC. I'm going to end the biggest JobServerChild.exe processes on our 4 clustered servers and see what effect it has.

[Former Member]

ending the rogue JobServerChild.exe processes on all four servers has apparently had no ill effect on the system and, fingers crossed, has fixed our slow running CMC problems. For now.

[Former Member]

I've done some more inversigation on this, and it seems to principally be affecting scheduled reports created in deski with multiple queries in the report. I've re-created identical reports in webi without getting the problems, so it's now less of a problem for us.

However, the above is very much a workaround rather than a solution, so if anyone does have further ideas on this I would be grateful.

[Former Member]

I faced the same issue of 100 % CPU usage and i kill all the jobserverchild.exe.But i dont know what is the impact of that,whether the scheduled jobs are also killed and if killed then how can we find which jobs are affected by this and how we can excatly find which exe is part of which scheduled jobs.

Can anyone please help on this.

[Former Member]

Any answer on this? I have this current issue. Is it the number of jobs scheduled at once?