dCache Tier I meeting November 27, 2018

[part of a series of meetings]

Present

dCache.org(Paul), IN2P3(), Sara(), Triumf(), BNL(), NDGF(), PIC(), KIT(Xavier), Fermi(), CERN()

Agenda

(see box on the other side)

Site reports

KIT

Upgrades

Last week downtime for LHCb was fine.

Everything went smoothly, finished down-time early.

No complaints so far, but these are early days.

The upgrade for LHCb took some 5--6 hours.

Plan downtime for ATLAS on 2018-12-11. That is the last production dCache instance upgraded to dCache v3.2.

ssh admin interface

No update on the ticket.

RT ticket 9418

Paul to remind Dmitry to update ticket with progress.

out-of-memory

RT ticket 9537

Paul and Al: who will analyse heap dumps.

Restart with wrong defaults

RT ticket 9545

Wrong limit reported

RT ticket 9516

dCache pools come up with 8 EiB with no limit.

We should somehow handle unlimited so we don't show the default value.

Tigran to investigate further.

zookeeper problem

RT 9542

ZooKeeper? cluster broke apart and pool stopped being well-known.

Xavier to add information about core domains (there being three, each running no services).

nearline archetype

Student develop nearline plugin for dCache -- wiki page on github repository.

Maven archetype -- this did not work.

Open issue in github or email support ticket.

NDGF

One question -- looking at switching from Oracle to OpenJDK 8. This is not available as a binary tar-ball release, but only available through the distribution.

What is the current status of dCache support for OpenJDK 11?

OpenJDK 11 is now part of our test environment -- current 'master' seems to work fine. We plan to back-port any necessary changes to dCache v4.2. Once this is done, we will look at dCache v3.2.

CentOS 7.6 will include OpenJDK 11.

NDGF tested running dCache v4.2 on OpenJDK 11. There were warnings logged, but it basically worked.

Support tickets for discussion

[Items are added here automagically]

DTNM

Same time, next week.