I second the LVM thing. I haven't done any benchmarks, but the added flexibility will make life a whole lot easier. If you had a 1 TB storage, and then you were making it 8 TB, with LVM, you can do it without your users ever knowing. Well, except for the fact that they will stop running out of space.
lvextend and pvextend Well, there was a change in the command structure. It has been a few months since I took a running system from 1 TB to 2 TB. But, it was easy. New disk in, rebooted it, add the disk to the logical volume group. Then, I just extended the partition size live. I don't know too much about number of inodes and the block size, so I guess that is where the tuning comes in. On Wed, Jul 29, 2009 at 03:44:32PM -0700, Ted Deppner wrote: > Cheap insurance against the need to change things. The only overhead > expense that would matter (because CPU and disk won't) would be your > time, differing between learning and using LVM versus having to change > something the hard way later on. > > On Wed, Jul 29, 2009 at 3:01 PM, Alex Mandel<[email protected]> > wrote: > > I think I'm missing why LVM would be good in this situation, it seems > > like extra overhead. RAID 6 in this case offers hot swapping of up to 2 > > failing drives at the same time, so I don't plan to do any mirroring > > beyond what the RAID config offers. As for the need to change > > partitions, aside from the OS which might have it's own drives, the rest > > of the drives are all 1 partition of 8 TB. > _______________________________________________ > vox-tech mailing list > [email protected] > http://lists.lugod.org/mailman/listinfo/vox-tech -- Brian Lavender http://www.brie.com/brian/ _______________________________________________ vox-tech mailing list [email protected] http://lists.lugod.org/mailman/listinfo/vox-tech
