Each of these flesystems were created on 1 physical disk each taking the whole space. ( The entrie disk has been taken up by the filesystems)
I wish to expand these filesystems; I have added five new disks to the system.
How do I expand these file systems without loosing the existing data?
I thought of the following;
Partition the new disks and create metadevices on them, ->create 3 volumes using SVM, -> mount the new volumes on 3 different fielsystems, -> then backup or copy the existing filesystems to the newly mounted ones (how do I do that?), -> edit the /etc/vfstab, change the mount devices for the filesystems so that the system mounts the newly created metadevices on the old mount points, re-partition the old disks and make them into metadevices, use growfs to expand the the 3 filesystems on the fly.
I haven't done this before, I just some opinions please. What steps can I take for these please
Do you like to use ZFS or would you take the "old fashioned" way ?.
With zfs you make a pool with your new disks an cut out your filesystems from the whole bunch.
Think about it. zfs could be the better way, possible the easier. SVM or ZFS ?
I don't know much about ZFS now; and I don't know if it can be implemented on this system without lossing the existing data. Then, will zfs co-exist with ufs? I forgot to mention that there are some other filesystems on this system that are already being managed by SVM.
I just read a brief about ZFS and it seems like it'll be the best, but it should have been implemented from the begining, right?
no,you can add zfs pools whenever you like.
The problem with your plan above is that you will have at the end a concated filessystem, no raid, no mirror, just a concat.
Yes I agree, maybe a concat filesystem for now as there's not enough disks for mirror. But there may be enough for raid. What do you suggest?
Maybe we can add a raid zpool on the new disks, create filesystems on the pool then copy the existing/old filesystems to the new ones on zfs (would that work? how?).