Event based hazard calculation blocks at 99%

[Paul Henshaw]

Running a an Event-based hazard calculation the user reports that the calculation proceeds normally until progress of 99%.
Subsequently one or two celery workers remain active at 100% CPU while the rest of the server cores are idle.
RAM and I/O usage are low, PostgreSQL processes are idle.
No errors are reported in the log.

This behaviour has been reported on two distinct servers.

The job file used refers to data which we cannot release publicly.

[Lars Butler (lars-butler)]

Is this still an issue? I know we have made a lot of changes/optimizations in the EB calculator in the last few months, so it's possible this issue has been fixed.