Hi folks,
what I usually do is creating the backup configuration files, downloading them to my local machine. Once this is done I'll strip the very first line containing the time stamp (this makes the files ready for the diff process in GIT, otherwise they will always change). Then I rename them so that they contain the identity (FQDN) of the MT device in your network (to make them unique) and commit them to a local GIT repository. Once this is done for the whole network (I usually use a text file with all FQDN of my MT devices) the commit is tagged and pushed to a private GIT repository.
Restore Configuration
Restoring MT configuration is done the opposite way using a version tag or commit ID. Remember to use a kind of restore text file so that units are initiated in the right order.
The whole process can be automated using a SSH key and the ssh agent. So the user has to enter the passphrase for the SSH key just once.
Collaboration
What about sharing the code in a Github repository? We could then further develop the solution so that it will be more robust and usable for more MT users?
Summary
- Download using SCP
- Strip the timestamp
- Continue with the rest of the network using a text file containing all the FQDN
- Commit to local GIT repository
- Create a tag
- Push the changes to a remote GIT repository