How To: Use Dump to back up a full filesystem
Way back in the day when tape drives first started being heavily used to do backups of Unix machines, the dump command was created. In typical Unix simplicity the dump command “dumps” files from one device to another device. This can be a tape drive, a hard drive, even a network share. rsync does a similar process but is meant for immediate use of those files.
First step is to be sure dump is installed. If not use, rpm, yum, port, apt-get, or your local repository method to install dump on your system.
The quickest command to get started is: dump -0 -j9 -f /pathtosavebackup /pathtobackup
This would give us a down and dirty dump of the requested path, or if / is used in the second part of the command, the full file system starting at the root.
-j9 tells the command to compress (using tar) the file as much as possible.
-f defines the device (or filesystem path) to dump too. Keep in mind that this command could not be used to dump files from the local filesystem back to the local filesystem.
The bad news is that this will take quite a while depending on how much data needs to be dumped, how fast the hard drives (or tape drive) is, and if backing up to a network share, how fast the ethernet connection is. In my tests 100gb filesystem got compressed down to 23gb, and took about fives hours across a 10/100 connection.
After the dump is done, tar can be used to unarchive the file to a new file system.