[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

[linux-lvm] LVM1 and software RAID1+0



Hi everyone. I recently setup a box running RedHat 8.0, which has LVM
1.0.3. The disks were 4xSCSI disks, /dev/sda - sdd. I got them setup in
a RAID1+0 setup, so

/dev/sda---|                              |---/dev/sdc
       raid1 /dev/md0---raid0---/dev/md1 raid1
/dev/sdb---|          /dev/md2            |---/dev/sdd

Ok so this was working great... I ran

pvcreate /dev/md2
vgcreate system /dev/md2
lvcreate .......

This all worked great, and I was able to mkfs on the lv's, and mount
them, use them, etc.

So I rebooted

vgscan, upon rebooting, refused to find any vg's. I ran 'pvscan' and it
showed /dev/md0 and /dev/md2 as  members of 'system'. [ding] the light
went on, I remembered that with LVM2 I had to create a device filter for
situations like this. 

So I searched, to no avail, for the same solution for LVM1. Could not
find it. The only way I've found around this is to wrap 'vgscan' with a
shell script that does this:

mv /dev/md0 /dev/md0.brb
vgscan $*
mv /dev/md0.brb /dev/md0

Which really sucks, and is totally hacky. It works, but I'm not happy
about it.

Is there anything I can do about this? I know LVM1, and especially 1.0.3
aren't exactly "current', but I still would like to see this documented
somewhere.

Thanks
-cb


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]