Last modified 2 years ago Last modified on 02/06/18 16:48:57

dCache Tier I meeting February 6, 2018

[part of a series of meetings]

Present, Tigran), IN2P3(), Sara(), Triumf(), BNL(Iris), NDGF(Christian), PIC(Elena), KIT(Xavier), Fermi(), CERN()


(see box on the other side)

Site reports


As Ulf has now left NDGF, the on-shift person will join the meeting. Christian joined today.

Active transfers

Christian mentioned that the active transfers page (webadmin) was "weird".

  • What is 'id' in the active transfers page?
  • The 'pool' column seems to be always empty.
  • Speed might have the wrong units (appears to be raw bytes-per-second, not kB/s).
  • Status has pool and mover id, despite these appearing in separate columns.
  • Preferred order is: host, pool, status.
  • What is "Proc"? Should be the pid of client -- but missing in almost all protocols.


KIT was working fine, but they are now struggling with ATLAS transfers from CERN.

ATLAS problems with FTS

Trouble with ATLAS trying to write with FTS.

Clients time out after 6 minutes, claiming no having received any performance markers.

Xavier to open a support ticket about this.

Paul suggested asking ATLAS what they have changed.


Last few days have this timeout issue.

FTS log "timeout" after 60 minutes.

They see lots of TCP connections on the door in CLOSE_WAIT state.

Some transfers are fast, some are slow.

dCache was *not* recently updated.

Restarting the door seems to help -- at least the CLOSE_WAIT state transfers went away.


Nothing special to report.

Will start planning the update to dCache v3.2

Support tickets for discussion

[Items are added here automagically]


Same time, next week.