dCache Tier I meeting October 27, 2009

[part of a series of meetings]

Present

dCache.org(Tigran, Irina, Owen, Jan, Tanya, Paul), IN2P3(), Sara(), Triumf(), BNL(Pedro), NDGF(Gerd), PIC(Gerard), GridKa(Silke, Doris), Fermi(), CERN()

Agenda

(see box on the other side)

Site reports

NDGF

Quite for a while now.

Moved most of the head nodes to new hardware. The dCacheDomain node is still on the old hardware (due to DNS issues with the pools).

Gerd mentioned on

PostGreSQL 8.4 dCache v1.9.5-5 upgrade

Tomorrow afternoon switching to new info provider.

Things looking.

Choice between

newer kernel to older

Hardware is sufficiently new that we would need to

NFS access for dCache? No, not for the WNs.

Use Chimera as name-space.

Gerard: we wish to switch to Chimera asap. We have some small VOs.

GridKa

Update today ATLAS to 1.9.5-4 update was fine; it's

Suffering massively from the root directory problem. Work-around is in place, but more than half of the directories created are root-owned. Gerd: The 1.9.5-5 released today contains some work-around

Only need to update the PnfsManager to get the benefit.

Don't believe we can migrate ATLAS from PNFS to Chimera before LHC start.

info system was a bit unstable earlier on.

Not sure if it was a dCache problem or something else.

New XML format was more complex. Tried to use the old XML file to begin with. Some more information was published but other was not.

CMS 1.9.4-2 Next week old dCache 1.9.1, Wednesday CMS 1.9.5-latest

PIC

There were a

File size in level -2, fixed in 1.9.5-5 in PnfsManager node. Issue with PhedEx.

Can we upgrade only machine with PnfsManager to 1.9.5-5? Yes.

New info provider: found something publish several SRM end-point.

The reason to publish

Multiple SRMs on your system.

We tried this once but had issues with GridFTP.

Need to have only one SrmSpaceManager in the whole dCache system.

With 1.9.6, this becomes easier because it is easier to move cells about.

BNL

• java.lang.ArrayIndexOutOfBoundsException on SRM 1.9.4-3 : ​http://www.dcache.org/rt/Ticket/Display.html?id=5202 RT#5202
• pin manager breaks when trying to remove pins which it cannot find in it's database (problem seen when using the migration module)` : ​http://www.dcache.org/rt/Ticket/Display.html?id=5196 RT#5196
• -allow-relay option is missing from the gridftpdoorSetup therefore on each upgrade one needs to remember and edit the gridftpdoor.batch file by hand : ​http://www.dcache.org/rt/Ticket/Display.html?id=5173 RT#5173
• need to use vorole-mapping since gPlazma with any other option is making SRM fail internally with role mismatch (!)

NFS

Testing of NFS. It's not totally straight-forward to test at the moment: you need a new kernel and new NFS client tools.

Gerard: we hope that, by the time enstore supports Chimera, there will be kernels and user-land tools to support NFS v4.1

Tigran believes that OpenSolaris? Build 125 doesn't support NFS v4.1.

DTNM

Proposed: same time, next week.