Using rsync to enable automatic backup MX mail server domains

[aarondwyer]

Hi

This is for the techies out there. So flex your techie muscle on this...

I have a redundant mail server setup and I've got a crontab setup to rsync copy my /etc/localdomains on the primary to /etc/secondarymx on the fallback mail server.

Like this from the backup mailserver crontab

0 1 * * * rsync -e ssh -avz root@primaryserver:/etc/localdomains /etc/secondarymx 2>&1 > /var/log/nightly_email_backup.log

However I'm noticing that the rsync takes a direct copy of the file from the primary server, but I thought it would just append by default.

I have domains that I've manually put into the /etc/secondarymx on the backup server that keep dissapearing after the crontab has run. So the rsync is removing them as well as adding in the new domains.

I just want the rsync to append the new domains that end up in the /etc/localdomains on the primary at the end of each day, whatever is in /etc/secondarymx can stay there even if the domains are removed from the primary.

Does anyone know how I can do this with rsync or scp or whatever.?

Thanks
Aaron

[jerrybell]

Seems like it might be as easy as creating a script that looks like this:
rsync -e ssh -avz root@primaryserver:/etc/localdomains /tmp/secondarymx
cat /etc/secondarymx >>/tmp/secondarymx
cat /tmp/secondarymx |sort |uniq >/tmp/secondarymx.new
mv /tmp/secondarymx.new /etc/secondarymx
rm /tmp/secondarymx



Then, put that script in cron and it should do what you want. Mine is probably not the only way to go, and there may be ways to optimize, but that's what I can pull out of my head at the moment.

[aarondwyer]

Thank you I will try that out tonight, the logic looks very sound, and I'm sure that would work so I'll report back my results.

Just a thought, doesn't rsync have an append function, would that work I wonder? I'll have to test that as well.

Thanks
Aaron

[jerrybell]

rsync does have an append fucntion, but it doesn't do what you think. Here is the man page excerpt:
--append
This causes rsync to update a file by appending data onto the
end of the file, which presumes that the data that already
exists on the receiving side is identical with the start of the
file on the sending side. If that is not true, the file will
fail the checksum test, and the resend will do a normal
--inplace update to correct the mismatched data. Only files on
the receiving side that are shorter than the corresponding file
on the sending side (as well as new files) are sent. Implies
--inplace, but does not conflict with --sparse (though the
--sparse option will be auto-disabled if a resend of the
already-existing data is required).

So, essentially, it is just another, more bandwidth efficient method to transfer the whole file.

Note: my rough script above would not remove a domain that was deleted from secondarymx on primaryhost. Otherwise, I think it'll work too.