Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Projects

  • NM Cluster scheduler replacementreplacement 
  • ALMA profiling
  • Alamo StarlinkStarlink  DONE
  • GPU support
  • NAASC scheduler replacement plan ngVLA Data Processing ConOpsPOSTPONED
  • VLASS, assist with cube imaging plan DONE
  • Create RADIAL specification

Discussion items

ItemWhoNotes
HERA hardwareJames
  • aocoss13 130466 racked, booted. Needs Lustre.  Stolen to repair aocoss04.
    • 2022-02-10 krowe: spoke with Matthew today and explained that HERA paid for a 13th OSS and there are only 12 in production.  He didn't understand the HERA angle but now seems to agree that the broken OSS should get fixed even if that means just buying a new chassis.  We will see.
    • 2022-03-16 there is motion on this.  I think CIS is replacing the hardware.
Slurm bugkrowe, jrobnett

Respond to SchedMD's question about suport contract.  Submitted a req (184349) for 1 year at $11,750.  PO: 375504

We now have a SchedMD account.  I need to submit some bugs.

CARTA glide-insJamesNeed to investigate gliding in based on lack of free slots rather than idle jobs.  Can one query HTCondor for a CARTA-shaped slot (core, mem, disk)?
Condor WeekAll

http://htcondor.org/HTCondorWeek2022 (May 23 to May 26)

Deadline for in-person registration is May. 2, 2022

  • ngVLA presentation?
  • nraorsync short presentation? ~krowe/nraorsync.odp
Vacation

krowe

fmadsen

Jun. 14 2022 - jun. 23, 2022 (7 days)

May 31 2022 - June 7 2022 (6 days)

NVMe Drivesjrobnett, kroweticket 139472 buy 30 NVMe drives for nmpost nodes.  What type to get?
Data reduction workshopfmadsenFrank Schinzel asked if I could support for a day to work to work with participants with imaging on the cluster - scheduled for October 18th.nraorsynckrowe

improved how stdout and stderr are copied back.  the _condor_stdout and _condor_stderr files no longer end up on the submit host but instead just what was specified in the submit description file.  This was a problem I had with CHTC and Todd last year.  I think he was right and I was wrong.  It happens.

But then I uncovered another problem.  If you don't set output and error, the plugin is never run on upload.  Damnit!

Jira
serverDMS JIRA
columnIdsissuekey,summary,updated,assignee,priority,status
columnskey,summary,updated,assignee,priority,status
maximumIssues20
jqlQueryfilter=12536 order by assignee
serverIdeb2e750b-a83a-387e-8345-36eee8a98f01

...