If the queue policies of the send-jobs queue are the same as the queue policies of the receive-jobs queue, the user should see identical behavior, whether the job is scheduled locally or remotely.
The job-level (user-specified) requirements and queue-level parameters (set by the administrator) are used to schedule and run the job.
If a job runs in the submission cluster, the send-jobs queue parameters apply. If a job becomes a MultiCluster job and runs in another cluster, the receive-jobs queue parameters apply.
Since the receive-jobs queue policies replace the send-jobs queue polices, LSF users might notice that identical jobs are subject to different scheduling policies, depending on whether or not the job becomes a MultiCluster job.
In general, queue-level policies set on the execution side are the only parameters that affect MultiCluster jobs:
If the job requirements conflict with the receive-jobs queue parameters, the job is rejected by the receive-jobs queue and returns to the submission cluster.
Runtime queue level parameters (terminate when, job starter, load threshold, exclusive, etc): the receive-jobs queue settings are enforced, the send-jobs queue settings are ignored.
Resource requirements: the receive-jobs queue settings are enforced, the send-jobs queue settings are ignored.
Resource limits: the execution cluster settings are enforced, the submission cluster settings are ignored.
Job slot limits (hjob limit, ujob limit, qjob limit): the execution cluster settings are enforced, the submission cluster settings are ignored.