Last modified 3 years ago Last modified on 10/14/14 14:45:09

dCache Tier I meeting October 14, 2014

[part of a series of meetings]

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


(see box on the other side)

Site reports


There was an issue with ATLAS SRM becoming overloaded due to activity from FTSv3 server, reported as RT 8475. The activity seems mostly with it attempting to create/upload many files that already exist. These operations all failed because site-local configuration disables overwrite operations.

The problem was that the SRM suffered from high CPU load. It wasn't clear whether this load was from establishing the TCP/SSL connection or whether it was due to internal activity within dCache. dCache log files contained messages indicating the database wasn't able to keep up with the load.

During the meeting, Tigran sent some PostgreSQL configuration that should allow the database to provide improved performance.

KIT's ATLAS instance is currently running dCache v2.6.33

dCache team to talk to FTS guys to ask what is happening.

Keep same connection -- if same user

Xavier said that it appears that everyone is now using FTSv3.

xroot issue

KIT have upgraded their dCache to fix the xrootd problem. They have received no feedback on whether this has fixed the problem.

They are also waiting for confirmation that the Brazilian CA problem is fixed.

Paul said he would drop Philippe an email to try to prompt a response.

Support tickets for discussion

[Items are added here automagically]


Same time, next week.