November Meeting notes 11-21

Date

01 Nov 2021

Item	Who	Notes
HERA hardware	James	herastore02 135532 herastore02 racked, powered, OSed, sotrcli, cards moved. Needs correct cards. Buy two 9380-8e cards Arrived Friday Nov. 12, 2021 Peter will install 02a racked, disked, firmwared, powered, SASed. Needs format, mount. aocoss13 130466 racked, booted. Needs Lustre. Stolen to repair aocoss04.
Order test GPUS	jrobnett	Tesla T4, SW, HH, 16.5cm long, in nmpost064 RTX A4000, SW, FH, 24cm long (too long for R640), in gpuhost003 RTX A5000, DW, FH, FL, in gpuhost003
Glideins	krowe	135553 Port RHEL-7.8.1.5 to CV. Started. cvpost006 135799 Install HTCondor on cvpost. Started, needs factory/pilot scripts.
T4 HERA	krowe	Working: Felipe is done with nmpost064 So I told plaplant about it.
MPI and Slurm	krowe	Understand and better document how to use MPI in Slurm
HTCondor plugin	krowe	fix _condor_stdout and _condo_stderr in the plugin if possible. Tell Charlotte
HERA GPU in Slurm	krowe	Working: Put a herapost GPU server in Slurm such that it requires a special something to use or at least is last in line to run jobs. Tell plaplant when ready to test.
Slurm	fmadsen	test some casa pipeline jobs with Slurm (parallel, multinode, machinefile, no machine file, no -n)
Sprint or CHTC	jrobnett	decide if we want to meet with SSA or CHTC at 1pm on Nov. 23

CAS-13342