Glasgow SRM Setup

From GridPP Wiki
Revision as of 21:05, 3 August 2007 by Graeme stewart (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

This page is intended for sites to use as a template in order for them to describe their SRM setup in detail. Obviously, sites can change/alter the format to suit their specific needs, but do not edit this page! The questions posed below are just to give sites an indication of the sort of information that should go here. Suggestions for additions are welcome and should be sent to Greig.

SRM Endpoint

 1.1.0: srm://svr018.gla.scotgrid.ac.uk:8443/
 2.2.0: srm://svr018.gla.scotgrid.ac.uk:8446/

Which VOs are supported?

alice
atlas
babar
biomed
camont
cdf
cms
dteam
dzero
ilc
lhcb
ngs
ops
pheno
sixt
supernemo.vo.eu-egee.org
totalep
zeus

Admin node(s)

DPM headnode is svr018.gla.scotgrid.ac.uk. Runs DPM, DPNS and SRM daemons (v1 and v2). Only a small filesystem on this host for testing purposes - usually set to read only.

Disk Server(s)

Nine grid disk servers, disk032-disk041 (but not disk037, which is for "local" cluster disk).

Hardware 22x500GB Hitachi SATA drives connected to an Areca PCI-X card.

Disk servers are setup as RAID 6 with 1 hot space, so have ~9.5TB of usable space.

To support the Areca card SLC is used (currently SLC4X, i386).

Optimisation

Filesystems are ext2 formatted. Will move to xfs once x86_64 is supported by DPM (see Performance and Tuning). Having no filesystems on headnode improves transfer reliability.

Additional Information

DPM has been a pretty trouble free experience.