The Instant Cloud allows you to offload optimization computations from one or more client programs onto a cluster of compute nodes. You can set a limit on the number of simultaneous jobs each compute node will run. When this limit has been reached, subsequent jobs will be queued. By default, the queue is handled in a Fist-In, First-Out (FIFO) fashion, unless job priorities are specified.
Queuing is not available for bronze licenses of Instant Cloud, and each pool created under this plan can only consist of one node with a job limit of one. Users under this plan may get the error message Job queueing feature is disabled when a client fails to dispose all models and environments at the end of the session. The article How do I release a shared license? explains how to avoid this issue.
- How do I resolve the error "No compatible runtime available for version X.X.X" ?
- What is the appropriate value for a Compute Server JobLimit?