SplineMPE scans fail to start

[mlincett]

Currently, scans with SplineMPE fail to start because the server hits the 8 GB memory limit.

This also causes an inconsistency in the database since the manifest can be retrieved from the scan_id but the scan_id does not appear in the list for that particular event.

Example: 7ebf7b83913948e2bd12d52703ecace9.

[ric-evans]

I'll add a REST argument to increase the server's memory request: #62

[ric-evans]

The data inconsistency is consistent with #57

[ric-evans]

Note that this was hitting a limit of 512MB, the client workers have a default of 8GB

[mlincett]

Note that this was hitting a limit of 512MB, the client workers have a default of 8GB

Thank you. SplineMPE does not load spline tables server-side but still loads some IceTray modules to pre-process the event. It makes sense that I would use more than 512 MB but I would not expect huge consumption nevertheless.