I have a production server running of a single disk, and want to add
a disk and use LVM mirroring to make all lv's RAID1. So now I've read
up a bit on LVM mirroring, and am a bit confused about the mirrorlog..
It seems to be recommended to use a third physical disk for the
mirrorlog, which seems silly, and is impossible for my situation as
the blade can only hold two disks.
One option is the to use the in-memory log (--corelog), but that
requires full resync on every boot, which also doesn't seem ideal.
So I wonder if it can be possible to create a mirrored log volume,
that itself has an in-memory log (--corelog), and specify that the
data volumes use this log volume for log. Then only the log
volume(s) would need to resync after a reboot, and that should be a
lot less data to sync.
i.e. something like:
# Make our volume group:
vgcreate rootvg /dev/sda1 /dev/sdb1
# Create mirrored log with in-memory log:
lvcreate -n mirrorlog --size 100M --corelog -m1 rootvg
# Create volume using mirrored log:
lvcreate -m1 --size 10G --mirrorlog disk rootvg
--logdevice=rootvg/mirrorlog
Did that make sense? Does anybody know how where the magic
--logdevice option is hiding, or how to achive something like this ?
-jf
_______________________________________________
rhelv5-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/rhelv5-list
