Configure a remote-only queue

To make a queue that only runs jobs in remote clusters, take the following steps:

  1. Edit the lsb.queues queue definition for the send-jobs queue.
    1. Define SNDJOBS_TO. This specifies that the queue can forward jobs to specified remote execution queues.
    2. Set HOSTS to none. This specifies that the queue uses no local hosts.
  2. Edit the lsb.queues queue definition for each receive-jobs queue.
    1. Define RCVJOBS_FROM. This specifies that the receive-jobs queue accepts jobs from the specified submission cluster.

Example

In cluster1:

Begin Queue
QUEUE_NAME = queue1
HOSTS = none
SNDJOBS_TO = queue2@cluster2
MAX_RSCHED_TIME = infinit
DESCRIPTION = A remote-only queue that sends jobs to cluster2.
End Queue

In cluster2:

Begin Queue
QUEUE_NAME = queue2
RCVJOBS_FROM = cluster1
DESCRIPTION = A queue that receives jobs from cluster1.
End Queue

Queue1 in cluster1 forwards all jobs to queue2 in cluster2.

Disable timeout in remote-only queues

A remote-only send-jobs queue that sends to only one receive-jobs queue.

Set MAX_RSCHED_TIME=infinit to maintain FCFS job order of MultiCluster jobs in the execution queue.

Otherwise, jobs that time out are rescheduled to the same execution queue, but they lose priority and position because they are treated as a new job submission.

In general, the timeout is helpful because it allows LSF to automatically shift a pending MultiCluster job to a better queue.

Force a job to run in a remote cluster

Jobs can be forced to run in a remote cluster at submission.

Use bsub -q and specify a remote-only MultiCluster queue.

This is not compatible with bsub -m; when your job is forwarded to a remote queue, you cannot specify the execution host by name.

Example:

queue1 is a remote-only MultiCluster queue.

% bsub -q queue1 myjob
Job <101> is submitted to queue <queue1>.

This job will be dispatched to a remote cluster.