[Halld-offline] bottle neck on /volatile

[Alexander Austregesilo]

Dear Colleagues,

The 2023-01 REST production is progressing well, we have already 
processed 16% of the data in the last 10 days.

The output files are copied to /volatile before they can be moved to 
tape and this transfer seems to be our current bottle neck. As a 
consequence, we are seeing high incoming data volumes on the /volatile 
disk. JLab computing operations kindly expanded our quota to 500TB, but 
we still fill up nearly 85% of it.

Could you please refrain from writing large competing data volumes to 
the volatile disk while we sort out this constraint? It would also be 
immensely helpful if you could delete files that are no longer needed or 
back them up to tape if you want to keep them. Owing to the large 
throughput, old data is in danger of being swept away by the deletion 
demon anyway.

You can check your own usage here:

https://scicomp.jlab.org/scicomp/volatileDisk/userUsage

Thank you for your help,

Alex and Igal

-- 
Alexander Austregesilo

Staff Scientist - Experimental Nuclear Physics
Thomas Jefferson National Accelerator Facility
Newport News, VA
aaustreg at jlab.org
(757) 269-6982