This is a discussion on Restoring LVM after a Hard Drive Failure (no raid) - Redhat ; At the moment, I am running Fedora 8 with 5 physical hard drives (all of different sizes) but *NOT* using LVM. I have a lot of data spread all over the place and it is clear that I need to ...
At the moment, I am running Fedora 8 with 5 physical hard drives (all of
different sizes) but *NOT* using LVM. I have a lot of data spread all over the
place and it is clear that I need to (1) back it all up more often and (2) more
efficiently use the space. So I am setting up a "backup server" using a spare
machine that will sync up to the main server each night. So I am thinking that
since the data will be duped anyway now might be a good time to actually think
about a strategy instead of just flinging drives at the machine every so often.
In reading about LVM it sounds intriguing, but I can't seem to find out what is
involved if a physical drive dies within the volume group. Lots of articles
seem to say you should use raid for redundancy but I pretty much will have a
mirror with the other machine.
If you don't use raid and one hard drive out of 5 in the LVM volume group dies,
is there no way to get the data off of the remaining 4 (ie. does the volume
group just get "corrupting" and that's it)?
My plan was to have the /, swap, and other system directories on 1 small (40gb)
drive outside of the LVM. Then put the other 4 drives (totalling 1.4 TB) into a
LVM group. The "backup server" would be similar with 1 drive out of the LVM for
the system, but 3 drives in an LVM group totalling 1.5 TB.
Wasted space issues aside, is this setting myself up for a major hassle if I
lose a hard drive (hopefully on only one machine at a time!!!)?
Can anyone think of a better solution (with pros and cons)?