DFSc Team re-organization, given Lori rise to CSCF RSG Manager, Nathan obtaining an RSG Research Support position and Anthony becoming FTE Infrastructure:
Anthony is now project lead for the DFSc service with Lori stepping back from the team
Nathan (RSG) and Guoxiang (Inf) being backup in-depth experts on DFSc and Cephfs
Clayton, Guoxiang and (Inf Empty Contract Position) being experts on GaneshaNFS service.
Nathan and (Inf Empty Contract Position) on NextCloud (Vault) service.
Clayton and (Inf Empty Contract Position) SAMBA CIFs access to DFSc
smbfiles.cs.uwaterloo.ca has 10 people using during this meeting
smbfiles.student.cs.uwaterloo.ca (it's ready week so we'll resource usage monitoring later)
People thoughts on performance since June 2023, specifically around NFS.
Is DFSc providing an acceptable minimal level of service via?
Should we accelerate DFSc evergreen cycle by one year this fiscal year?
Result of discussion is that yes, purhasing the new storage nodes with U.3 drives ASAP is the way to go.
items that were discussed:
This would be via the addition of three 24 2.5" bay NVME storage nodes with ten 12.6TB mixed use U.3 drives to this fiscal year to add 100+TB raw storage.
This would mean delaying getting a dedicated OOB CSCF Admin cloud node and the budget line items to replace 6TB drives per 421 nodes with 18TB drives to a future fiscal year as well a a large piece of our disk replacement supplies budget.
DFSc priority RT items:
RT #1304328: Deploy modified patch to prevent NULL ptr dereference in kernel crashing CephFS clients (Found by Guoxiang)
Information in this area is meant for use by CSCF staff and is not official documentation, but anybody who is interested is welcome to use it if they find it useful.