Hello, I'm in this situation and would like to be certain before doing bad things... Thanks in advance for your help rh el 6.1 host and a vm with one disk that is LVM based (actually it should apply with recent upstream kernel and lvm too) - shutdown guest - lock exclusively the lv (because it is actually a cluster of two rh el 6.1) on one node: # lvchange -an /dev/VG_VIRT01/zensrv_002 on the other one: # lvchange -aey /dev/VG_VIRT01/zensrv_002 - create snapshot # lvcreate --size 5G --snapshot --name zensrv_002_snap /dev/VG_VIRT01/zensrv_002 Logical volume "zensrv_002_snap" created - power on vm going ahead doing some intrusive tests Note that I maintained the original LV as the disk for the VM - At the end of my work the situation I have changed about 1.7Gb of my original disk: # lvs /dev/VG_VIRT01/zensrv_002* LV VG Attr LSize Origin Snap% Move Log Copy% Convert zensrv_002 VG_VIRT01 owi-ao 10.00g zensrv_002_snap VG_VIRT01 swi-a- 5.00g zensrv_002 34.77 - Now I'm confident with the changes.. Is it ok to simply remove the snapshot? - What if I'm not confident with the changes and would like to revert to the original state ? Will it do the command lvconvert --merge /dev/VG_VIRT01/zensrv_002_snap Or did I make it totally wrong and had to work with the snapshot and not with the original LV???? Keep on reading docs, both rhel and general but I'm sort of confused.... Probably someone on this list already went through this way... And could I have been working with both the LVs at the sane time (for example creating a new vm base on the snapshot lv) and what would be the possibile scenarios in that case? Gianluca -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
