2009/4/28 James Bensley <jwbensley@xxxxxxxxx>
Linux software raid1(e) allows any number of drives to be used such that each block exists on two devices (i.e block 0 on drives a and b, block 1 on drives c and a, block 2 on drives b and c etc...), and it's possible to grow an existing array onto additional drives, though it could take some time as the data gets redistributed across the disks.
His fdisk output shows that software raid is being used such that the filesystem exists directly on top of the md0 device...
d
Hi there,
Just to clarify; You have two 1TB disk that are mirrored (raid 1)?
Normally a mirror can't be extended as the mirror is between block 0
to 1000000 (for example if 1000000 was the end of the drive) on one
drive with block 0 to 1000000 on the other (if you have used block
level mirror, you could of chosen byte level mirroring?). With RAID 1,
mirrors require an even number of disks.
Linux software raid1(e) allows any number of drives to be used such that each block exists on two devices (i.e block 0 on drives a and b, block 1 on drives c and a, block 2 on drives b and c etc...), and it's possible to grow an existing array onto additional drives, though it could take some time as the data gets redistributed across the disks.
How is you RAID set up, is it software or hardware? If hardware, what
raid controller/card are you using? How much are you looking to extend
the storage capacity by? How much does redundancy matter to you? Do
you just want speed (i.e. a stripe, RAID 0)?
His fdisk output shows that software raid is being used such that the filesystem exists directly on top of the md0 device...
d
_______________________________________________ CentOS mailing list CentOS@xxxxxxxxxx http://lists.centos.org/mailman/listinfo/centos
