Is a 2TB disk, shared with 2 servers.
You mean a cluster with 2 nodes? Can you use OCFS2 for the purpose to share disk space between 2 servers without running a cluster?
If you run a cluster you can't just be using a local hard drive. You need some storage controller that allows to share disk space. Most modern storage controllers allow to manage allocated storage space dynamically and increase the size of partitions or LUN's. Unless you use some conventional RAID solution which requires to re-initialize the storage.
Creating OCFS2 volumes on logical volumes (LVM) is not supported, but why LVM, instead of resizing the storage at your storage controller? You can resize an OCFS2 volume using "tunefs.ocfs2 -S", but the volume must not be mounted by any node in the cluster.
Yes, it's a storage LUN shared between a two servers cluster using OCFS2 filesystem. So, my two servers can see same files placed there.
Based in your last paragraph, I need to unmount OCFS2 filesystem and them go to my Hitachi storage and resize the lun from 2TB to 4TB. Then, i need to run "tunefs.ocfs2 -S /dev/xxxx".
Thanks for your patiente.
Yep, as far as I know. As always, make a backup first.
I will try it and report here the results.
Here's the end of story.
Our OCFS2 filesytem is mounted over a single 2TB Hitachi storage LUN.
The LUN wich needed to be expanded from 2TB to 4TB were located in a storage disk pool which had no more space, so I can't resize it.
Well... we did a backup of LUN data, unmount OCFS2 filesystem and (via hitachi software) we delete the LUN paths. Also we delete LUN multipath identifiers (multipath -f mpathXX).
Then we choose another storage disk pool with sufficient space (4TB) and we create a new 4TB LUN. Map it to server and rebuild OCFS2 over this new LUN.
Thanks for tips DUDE!