wiki:developers-meeting-20100511
Last modified 11 years ago Last modified on 05/11/10 16:34:32

dCache Tier I meeting May 11, 2010

[part of a series of meetings]

Present

dCache.org(), IN2P3(), Sara(), Triumf(), BNL(Pedro), NDGF(Gerd), PIC(Gerard), GridKa(Doris), Fermi(), CERN(Andrea)

Agenda

(see box on the other side)

Site reports

GridKa

Everything fine; opened a ticket last week about. How to delete some files from PNFS.

Running smoothly

NDGF

One problem discovered yesterday ..

some files we tried to stage from tape on behalf of ATLAS. Some reason the stage never completed: always on-line. Files stays on stage-pools; normally the files were pinned on stage pool. Don't know why the file was pinned on the stage pool. Some reason Pool Manager included stage pool as a possible

Otherwise things are fine

PIC

Running 1.9.5-17 everything OK.

Running 1.9.6: having difficulty with WebDAV.

Did you get it work with Gnome or KDE?

I can browse with Firefox. Problem was the client gets 302 redirect but

(is 307 the better redirect message). Next major release 1.9.8 will have proxy for reads too:

CERN

Nothing

Talk tomorrow GDB about data-management .

BNL

Everything is working well with the

Things Smooth.

dCap timeout

Ticket a few days ago.

Does it happen on close?

Yes, I believe the WN can read the file completely but then they loose connection.

Server-side, we see IllegalRuntimeException?

The client always retry: after five attempts

Client looses TCP connection.

Do they continue to read? No, I believe the client has already read the data.

"Last IO operation." This just means the last operation failed.

Tigran will verify the log files

Does it happen often? 5% of all jobs.

Usually starts to appear when we start to saturate the network.

Maybe need to update the dcap library on one of the worker-node machines. Problem

Support tickets for discussion

[Items are added here automagically]

DTNM

Proposed: same time, next week.