When submitting jobs on a *real* CXC system,
PBSP will NOT set a limit on the Cray Compute nodes.
In fact, apRRn will be responsible for the setting up of the job environment on the CCCnode.
Therefore, memory management will be enforced and managed by around CCDDD.
WRT the qsub wrapper issues, Chiba is estimating all issue reported as of today (Dec-02) will be addressed by Dec-09.
Please note that we do NOT consider issues #19, #21, and #22 as issues.
Please review our comment in XLS.
Although there are issues found with qsub wrapper,
the users can submit jobs to the cluster.
IOW, it is NOT broken.
Are the issues you are reporting show-stopper from deploying on Dec-06?
WRT planning for Dec-06,
if the node allocator review meeting is positive and the qsub wrapper issues reported are NOT show-stoppers,
then we can plan to update KyotoU subsystem A on Dec-06.
We look forward to your guidance.