Seek help from all experts,
I understand that if I perform a savevg on a VG, and later perform a restvg to another hdisk, I will get back a exact same VG that I perform savevg earlier. It will be identical even on permission bit, owner and group.
However, if I perform a restorevgfiles for a filesystem from the same savevg file, will I get a identical filesystem? In terms of permission and owner and group will be identical as well?
Firstly, the restore would need to be to a server that does not have this volume group. I don't think you can clone to the same server because there would be naming conflicts.
The restvg command requires a suitable number of PVs to restore to and these are completely consumed by the process, although that might mean there is free space within the recovered volume group.
To get us to a known point, can you confirm how you saved the volume group (did you exclude any files?) along with any information about the original you have, e.g. number & size of logical volumes and if they are mirrored, PP size,
We would then need to know the target disks and if there are any naming conflicts for the target server. The PV names to restore to can differ from the original. We will need to know quite a lot to work out how to help.
Kind regards,
Robin
Not exactly. You don't need to have exactly the same physical volumes (i.e. they need to be big enough to hold all the LVs but might be bigger than their originals), etc.. They just need to be able to accomodate the VGs needs. That means: if you have i.e. LVs with copies and "strict" as a rule you need two disks so that the copies can go on different disks and similar.
savevg is using the backup command to save the files themselves (this backup image is part of the savevg image and if you know what you are doing you can even pull out single files from it using the restore command).
In fact savevg works exactly the same way as mksysb (minus the boot code, etc., because the base mksysb image is created by doing a savevg rootvg ) which we have described several times in this forum.
I hope this helps.
bakunin
As bakunin mentions the commands mksysb and savevg are scripts that act as "front-ends" and are able to collect and record information needed to recreate an "identical" copy during restore.
a) it is not advisable to restvg a volume group that you have saved, but not removed. While it may be possible to provide an argument to give the volume group a new name, by default restvg wants to re-create the same logical volume names. As the original volume group exists - this is not acceptable and AIX will rename to logical volumes and your "restore" is no longer identical at the volume group level.
b) as to restore files from a "bff" (backup file format) file - these files will have identical information in the inodes - again, at a binary level - not necessarily at a user level. If UID and GID (i.e., /etc/password and /etc/group) are identical on "both" servers, the restore will appear identical, if the USER/GROUP information is not identical - the restored files will not appear as identical. At a binary level - they are identical.
In short, when you need identical information at an INODE level (i.e, OS specific knowledge about special files and/or ACL, permission bits (e.g., tcb-bit) use the commands backup and restore for saving files, not applications such as ftp, scp, cpio, tar, rsync as they do not copy inode data at an OS level. They might work for an AIX to AIX copy, but there is no guarantee.
p.s. - if you goal is to copy a filesystem, e.g., through a pipe - feel free to look at a script I used (I generally just type:)
# find . | backup -if - | (cd /some/where/new; restore -xqf -)
Look at: http://www.rootvg.net/content/view/301/309/ (from 2009!)