GPFS has a facility for replicating data (mirroring), but in their
documentation they suggest that you use RAID as your primary redundancy
mechanism (at least in this copy of the docs):
http://publib.boulder.ibm.com/infocenter/clresctr/vxrx/index.jsp?topic=/com.ibm.cluster.gpfs.doc/gpfs31/bl1ins1122.html
Write speed obviously depends highly on the disks in your system, the
speed of your nodes, the speed if your network connections, and the
quality of your switch.
Regards,
Rob
Erich Weiler wrote:
There is no in-built redundancy yet. Unless you setup something like
drbd.
Hope you find this information useful.
Do keep us posted.
Hmmm... Having a failover to shared storage would be OK but would
involve procuring 80TB of shared storage... Of course I could use
node1-100 for filesystem1 and node101-200 for filesystem2 (some kind of
drbd implementation). Weighing this against IBM's gpfs, which ain't
free but they have a bit better redundancy I think... Any notes on
performance? Has anyone tested write speed of say, 3TB or something?
-erich
_______________________________________________
Pvfs2-users mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-users
_______________________________________________
Pvfs2-users mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-users
