Jobs erroring every 5:15

8.12/8.97.2.0

Queue with two "threads".

First job goes into processing w/ pid 3132 (Runbatch)
Second job goes into processing w/ pid 3220 (Runbatch)
Third job goes into waiting w/ pid 8628 (UBE Kernel)
Third job fails w/ pid 8628 (UBE Kernel) at a point in time that is 5:15 from some point.


Doesn't appear to fail *after* 5 minutes but *every* 5 minutes

Actually, every 5 minutes 15 seconds

Any job that is in 'W' status gets changed to 'E' status every 5:15.....like clockwork.

Hi,

I remember having seen a JDE.INI parameter (on E812) that
says how long should a kernel run before being killed.
It's kind of an expiration or auto-bounce parameter.

Kernel Recycling. Thanks but not the issue.


[ QUOTE ]
Hi,

I remember having seen a JDE.INI parameter (on E812) that
says how long should a kernel run before being killed.
It's kind of an expiration or auto-bounce parameter.

[/ QUOTE ]

It appears that because the MS cluster configuration has no dependencies for E1 Service cluster resource -> Disk cluster resource or E1 Network Name cluster resource the kernel(s) did not start properly.

Starting the E1 service cluster resource after the other cluster resources removes the problem.

Adding the proper cluster resource dependencies will solve the issue.


[ QUOTE ]
8.12/8.97.2.0

Queue with two "threads".

First job goes into processing w/ pid 3132 (Runbatch)
Second job goes into processing w/ pid 3220 (Runbatch)
Third job goes into waiting w/ pid 8628 (UBE Kernel)
Third job fails w/ pid 8628 (UBE Kernel) at a point in time that is 5:15 from some point.


Doesn't appear to fail *after* 5 minutes but *every* 5 minutes

Actually, every 5 minutes 15 seconds

Any job that is in 'W' status gets changed to 'E' status every 5:15.....like clockwork.

[/ QUOTE ]