On Tue, 17 Nov 2009 12:07:00 +0000, Steven Whitehouse wrote > Hi, > > On Tue, 2009-11-17 at 12:53 +0100, brem belguebli wrote: > > I think the constraint is just like for regular filesystems. > > > > 1 GB should be right, shouldn't it ? > > > Well there are journals of 128M each (default) so for two nodes, thats > 256M, so with one or two smaller structures (they take only a few blocks > each) 1G would seem a sensible minimum in this case, > Yes 128M is by default but the minimum is 8M, so 8x2 jurnals = 16M + some space for resource groups = 19M fs is possible with ~1M for data ... just confirmed: dd if=/dev/zero of=/tmp/test_gfs bs=1M count=19 losetup /dev/loop0 /tmp/test_gfs mkfs.gfs2 /dev/loop0 -j 2 -J 8 -p lock_dlm -t a:b mount -o lockproto=lock_nolock /dev/loop0 /mnt/tmp/ df -h result /dev/loop0 19M 19M 840K 96% /mnt/tmp > Steve. > > -- > Linux-cluster mailing list > Linux-cluster@xxxxxxxxxx > https://www.redhat.com/mailman/listinfo/linux-cluster -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
