Last modified 4 years ago Last modified on 10/10/17 16:09:18

dCache Tier I meeting October 10, 2013

[part of a series of meetings]

Present, Paul), IN2P3(), Sara(), Triumf(), BNL(), NDGF(Ulf), PIC(Marc), KIT(Xavier), Fermi(), CERN()


(see box on the other side)

Site reports


Ulf reported that things are currently going fine. There were some hiccups with tape over the weekend, but these were not dCache related.

NDGF is currently planning to upgrade their test system to dCache v3.2. They will also discuss with Mattias when to upgrade the production system to dCache v3.2.


After upgrading the ATLAS instance to dCache 2.16, Xavier has opened two new tickets: 9276 and 9277.


This reports a misleading error message -- should be easy to fix.


This ticket reports two potentially related problems.

First, srmcp is failing sometimes. There is insufficient information in the ticket to understand this problem -- we gave feedback on how to provide more information next time.

Second, a pool seems to have dropped out of dCache cell communication.

Use 'route' and 'ps' commands in System cell to help discover what is going on.

DB manipulation

Upgrade for ATLAS -- wanted to parallelise the DB update. Liquibase holds a lock.

Need a procedure on how to split cells into separate databases.

Advice given: run a fresh dCache and see which tables are created.

pool start-up time

ATLAS pools takes ages to start up.

Marc had mentioned: possible tweaking option je.log.filecachesize how many files files. Default is 100, but pool has 800--900 metadata files.

Tigran: Try to cat meta files to see if that helps. This warms the filesystem cache.

Xavier: will try this at some point.


Everything is OK.

Latest version much appreciated.

dCache and puppet

Final task at PIC is to migration to puppet 4. Want to use the repository dCache puppet repo, but who is using it?

Nobody yet -- feel free to modify.

Support tickets for discussion

[Items are added here automagically]


Same time, next week.