On Jan 16, 2010, at 11:46 AM, Les Mikesell lesmikesell@gmail.com wrote:
Adrian Sevcenco wrote:
Ross Walker wrote:
On Jan 16, 2010, at 10:27 AM, Adrian Sevcenco Adrian.Sevcenco@cern.ch wrote:
Hi! I have 2 identical hdd (1002FBYS) with the same firmware, etc .. really identical! when i try to add the second hdd to this raid1 array i have the message that /dev/sdb is not large enough to join array!! on the first hdd i have 100 mb unpartitioned at the end of disk (as i understood that mdadm put there some info...) Anyone, any idea what can be wrong? i already done this procees several times and this is first time i encounter this and i have no idea what can be the problem ..
Make sure the partitions are exactly the same, you can use sfdisk to copy the table from a to b.
well, i was doing raid over device (md_dX) but i found out the problem : even if there are "identical" devices, they have different number of blocks (hdparm output). how is this possible? is there a way to restrict the number of blocks that a hdd have? alternative would be just to clone the first hdd on the "smaller" one and then add the big one to and raid1 array ...
Did someone seen this situation?
I've seen a case where single drive volumes initialized in different (but identical) IBM 3550's with adaptec raid controlers would end up with slightly different sizes and wouldn't match up if moved. I think this was a bios or firmware difference but I'm not exactly sure. I was trying to clone disks and just started over with the smaller source so it would work everywhere.
Les might be on to something here I remember some BIOS store a per- device LBA setting which might cause to sizes to differ.
This brings up a good reason to use partitions as it will allow you to coerce the size down a little 128-256MB so if you need to replace a drive with a different manufacturer's drive in the future you will have a better chance in making it fit.
-Ross