dCache Tier I meeting MONTH DATE, 2011

[part of a series of meetings]

Present

dCache.org(Tigran, Antje, Tanja, Paul), IN2P3(), Sara(), Triumf(Simon), BNL(), NDGF(Gerd), PIC(), GridKa(), Fermi(), CERN()

Agenda

(see box on the other side)

Site reports

Triumf

Production was OK last week.

Two tickets: one is mostly clear, the other needs some more work.

Paul reported that DESY have been able to reproduce the problem, so Simon doesn't need to do more work on this ticket.

NDGF

Quite fine; nothing to report.

GridKa

Doris reported via email

One of our dCache instances is a bit unstable. We have trouble with the 
pools of one experiment, as a result the pools are offline. And it seems 
that these pools cause trouble on Pnfs (we had to restart pnfsd 
yesterday). Pnfsd was completely stuck and did not respond any longer. 
Today we had to restart PinManager and dCacheDomain. dCacheDomain run 
out of Memory, most likely the same problem we had some time ago, the 
pin manger tries to get a pin on files available only on offline pools.

Our other two dCache instances for CMS and ATLAS are working fine. Other 
than that, nothing to report from Gridka.

PIC

Gerard reported via email

We have an Enstore related issue with our Linux (CentOS5.5/RH5.6) pool
servers. We're in touch with Jon Bakken about it. This issue produces a few
kernel panic per month, this is what corrupted the berkeley DB in a pool a
couple weeks ago.

Besides this we've no issues. We will upgrade from 1.9.10 to 1.9.12 on our
Tier3 instance in a few weeks (need to find a maintenance window).

Support tickets for discussion

[Items are added here automagically]

DTNM

No meeting next week because of the EGI User-Forum and EMI Technical-Forum co-located meetings. However, support is available through the usual channels (support@…)

The next Tier-1 support meeting will be in two weeks time; Tuesday 19th April at the usual time.