This is a question to all GlusterFS User with big mdadm or zfs Brick. I was wondering that you have to stop the GlusterFS Volume if you want to increase the size of a Brick when you use lvm as Block device layer for GlusterFS according to the latest Redhat docu. In my opinion, the big advantage of lvm was to Resize Storagesize in Live Mode. Can anybody explain why Redhat do recommend this in this way?
Because i would combine GlusterFS with BTRFS I made a test with the setup: qcow2-File with btrfs and GlusterFS ontop. Then I start a VM on this storage and let it write random data to the vm image. Meanwhile I resize the btrfs volume +1G. The GlusterFS Daemon noticed the change instantly. No interruption in the vm I noticed. So I was skeptical. I made I new test with lvm and xfs as described in the redhat docu.
And the same here. No problems with Resizing Brick while writing to the Storage. What is the problem redhat has?
Because i would combine GlusterFS with BTRFS I made a test with the setup: qcow2-File with btrfs and GlusterFS ontop. Then I start a VM on this storage and let it write random data to the vm image. Meanwhile I resize the btrfs volume +1G. The GlusterFS Daemon noticed the change instantly. No interruption in the vm I noticed. So I was skeptical. I made I new test with lvm and xfs as described in the redhat docu.
And the same here. No problems with Resizing Brick while writing to the Storage. What is the problem redhat has?
Last edited: