Other Norton Products

[redk9258]

You could create a batch file to run after the backup is created and use RoboCopy to copy the files to the NAS.

[andreash_utah]

To take what Red said a step further, you can create a folder on your off-site NAS called 'archive'.  You can create a command file or use Robocopy to move or copy and overwrite the offsite files into the 'archive' folder. Ghost will ignore that folder. During the next offsite, when Ghost sees no files, it will sync it up again. Next backup, it starts all over. Just make sure you monitor the space usage or your NAS will fill up quickly.

[rgw]

Copying the files with robocopy would avoid the problem of Ghost thinking they are part of the recovery point set, so the files would not be deleted when the recovery point set was deleted (presumably).  But there is no easy way I know of to let robocopy know the name of the file being generated.  Robocopy would have to use wildcards (*.*) and although robocopy can skip files on the destination with the same size and timestamp, the destination is a Linux server running Samba, so I need to verify that the destination size and timestamp are interpreted to be the same as Windows.  However, this sounds to be worth a try, so I'll let you know.

[rgw]

To Andy - That's essentially what I'm doing only I don't use Windows Robocopy, I use the Linux 'ln' command to create hard links of the current offsite recovery point files into another directory.  So I'm using a command like:

   ssh -l root mybooklive 'cd /DataVolume/shares/myname/Norton\ Backups ; ln my-PC* ./recovery-backups'

This creates a "hard link" to each of the recovery point files in the "Norton Backups" directory in the subdirectory "recovery-backups".  When Ghost deletes the file entry(s) in the Norton Backup directory as a result of the original recovery point being deleted from the main backup directory (on the source computer), the hard links remain and are hidden from Ghost.  Because they are hard links, the file contents remain.  It has the same effect as copying the file to the subdirectory except that the contents are not duplicated and so it does not require twice the storage.  The reason this doesn't work exactly as I would like is that Ghost runs this command BEFORE the offsite copy is performed, so the link to the most recent backup (full or incremental) doesn't get made until the next time Ghost runs the backup task.  That's not a really big deal since I will only be missing the last member of the recovery point set on the NAS until the next time the backup runs, and I can always run the command from a Windows command window to update the links.

[andreash_utah]

That's an interesting approach. I am not a Linux guru by any means, but that does sound like it would work. Although be advised, our QA group has not tested that approach, so if it does not work, it may take some additional fancy scripting.

I am wondering if this contributes to the large incremental issue. Is it possible to disable this process for a few backups to see if it is?

[rgw]

Using hard links is a method available for Linux consisting of a set of Perl scripts called 'rsnapshot' combined with a copy program called 'rsync' to create multi-version file backups.  'rsync' is very robust and has features similar to robocopy but can also handle special file types on Linux such that the equivalent of the full system backup can be performed.  I've used rsnapshot to make backups and then used rsync to restore a complete bootable file system to a new disk for a Linux system.  'rsnapshot' uses hard links to make a "copy" of all files that haven't changed and uses 'rsync' to copy any file that has changed on the source.  This creates the equivalent of incrementals since the only space needed on the destination is that for the changed files.  I would have used this in my case to create the "backups" of the offsite copies except that the "first" time rsnapshot is run, it creates a physical copy of ALL files, so that would have doubled the space used on the offsite disk.  I consider using rsync just to perform separate file backups but that requires that an rsync server be installed on the NAS, and adding software to the NAS makes it more problematic when performing "firmware" updates.

Both rsnapshot and rsync are open source if you wish to take a closer look at their implementations.  Both can be run on Windows under Cygwin.

[rgw]

To add to my previous post, I discovered that the Western Digital MyBook Live NAS already has rsync installed and configured to run as a tcp or udp service.  Since I already have rsync installed on Windows under Cygwin, it can be used for file backups.  However, rsync does not compress the files, Ghost is still to be preferred for full system backups.  For file backups, I need to do a more thorough comparison of features and performance between rsync and robocopy.

[andreash_utah]

Ghost will also do file and folder backups in addition to full volume backups. Because I have not used rsync, I don't know how it compared.

[rgw]

Ghost file backups work okay but put everything into proprietary format, and the inferface for restoring files from a File Backup seems very awkward to me, i.e., there is no "browse" capability.  Sometimes I need to restore a file or folder that's been deleted, and with Ghost file backups, I have to search by name and/or wildcards.  I can use explorer type browse on recovery points, but that can be very slow.  So I prefer "plain" file backups with the same directory structure and file attributes that I can browse from Explorer or from the command line.  Not practical for the whole disk but works fine for selected directories, e.g., iTunes.  Just a personal preference.

[redk9258]

---

rgw wrote:

Ghost file backups work okay but put everything into proprietary format, and the inference for restoring files from a File Backup seems very awkward to me, i.e., there is no "browse" capability.  Sometimes I need to restore a file or folder that's been deleted, and with Ghost file backups, I have to search by name and/or wildcards.  I can use explorer type browse on recovery points, but that can be very slow.  So I prefer "plain" file backups with the same directory structure and file attributes that I can browse from Explorer or from the command line.  Not practical for the whole disk but works fine for selected directories, e.g., iTunes.  Just a personal preference.

---

I think File and Folder Backup sucks compared to a Recovery Point. Recovery Points are much easier and faster to find what you are looking for either by mounting the image as a drive letter or using recovery Point Browser.