dCache Tier I meeting July 18, 2017

[part of a series of meetings]

Present

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

Agenda

(see box on the other side)

Site reports

KIT

Assume everything is fine, based on zero complaints.

RT ticket 9071

A lot of warnings in dcap log file -- Ticket 9071

​http://rt.dcache.org/Ticket/Display.html?id=9071

PIC reported seeing the same problem.

Updated ticket with work-around.

PIC

Everything is running smoothly.

Staging stress test

This week CMS is working with PIC to stress their staging with a view to improve the performance. They have already doubled their performance.

They want to add lots of concurrent movers, to give Enstore the opportunity to reorder requests and avoid tape remounts.

What resources does the pool use per stage request:

The pool uses 3 threads per script invocation -- plus also the script invocation.

There's also the return code 72: if a script returns this value then pool will consider the request as having started and that the tape system is working on the request in the background. It will periodically check whether it has completed by issuing the same command.

NDGF

Unfortunately, NDGF was unable to connect.

Support tickets for discussion

[Items are added here automagically]

DTNM

Same time, next week.