I have BE 9.1 installed on our vcb proxy, and I am performing file level backups, full once a week and incremental daily. I understand the process to restore the full backup consists of:
-restore all the files created by vcbMounter to a directory, in my case, I am using a shared directory on the proxy
-share this folder to the ESX Server
-use vcbRestore against the files from the service console
However, how do my incremental backups come in to play? Restore each incremental backup to the proxy and then dropping then on the appropriate drive of the VM sounds a little crude. Does anyone have a best practice for this type of restore process?
I would go that route if I were performing image level backups, but I have been working with full file level backups and daily incremental backups. If I were to perform image level backups, I would have to do that daily, would I not?
What I should have asked is if it is possible to do incremental, daily, backups on an image level with BE 9.1.
The image is just that, an image, so you can't really do an incremental on the image itself. The image file will be seen as changed, which needs to be backed up, as opposed to incremental files being changed only requiring those to be backed up.
I suppose I should rephrase my question to what is a standard backup routine. Take image level backups once a week, and then file level backups daily incase of a deleted/corrupted file? If I went that route, I would think I would still need to perform a full file level backup at the beginning of the week so that the incremental, daily, backups are not backing up the entire VM. Restoring a complete VM would have an issue too. I could restore the image of the VM, but what about the files that have changed since that backup was taken? Would I need to use my daily file backups to replace outdated files?
The BIG thing to remember is that you need to be able to recover your systems in the event of a disaster. File level backups will let you recover the files. File level backups do NOT include the system state, so you still need this. Someone suggested using NTBackup in the pre-freeze script to catch the system state, and that is a reasonable idea, but it will take many steps to recover the systems.
In the BCDR plan and design engagements I have managed, I usually recommend at least weekly FullVM backups and then follow up with daily file levels. But this still doesn't get a fresh copy of the system state and you really need to consider that, if the system state changes frequently. So, for file servers, I use a hybrid method. Put the files on a separate VMDK. Back up the C: drive using the FullVM method and back up the data using the File level method. Transactional systems, like Active Directory, SQL and Exchange should really be backed up with a client daily and have a weekly FullVM of the C:.
VMware vExpert 2009
Careful. We don't want to learn from this.
Bill Watterson, "Calvin and Hobbes"