Ya, BS. I keep changing stuff. I chalk it up to not knowing what I’m doing combined with 4 A.M. thoughts lying in bed. But one was so damn critical…I actually did it before getting caffeinated this morning.
And I thought I didn’t know WTF I was doing…
So you can go read a previous post about the backup script or I can sum it up. You choose, I’ll just type:
- make full backup of my sql database in to www root
- tar nginx-conf file to www root
- tar-gzip www root directory
- upload tar.gz to home server
It occurred to me that steps two and three were a problem. My WWW directory will get big; in fact it’s quadrupled in size already. There’s a lot of media files and a lot of junk I put up there. Doing a complete full transfer, while redundant; is not the way to do it. So why don’t I do my hacky backups the right way:
#!/bin/bash mysqldump -u root --all-databases > /var/www/full.sql tar -czf /var/www/etc-nginx.tar.gz /etc/nginx rsync -arvz -e 'ssh -p <port>' /var/www user@server:/media/to/backup rm /var/www/etc-nginx.tar rm /var/www/full.sql
Rather than tar /var/www; it is literally better in every way to rsync it over SSH. This naturally avoids copying every file every single time…it’s only going to copy files that have been added or changed; exactly what we need. But it also allows me to save some bandwidth and copy large files in to the backup directory where they should already be; allowing me to pre-sync large files.
I already generated the initial rsync while debugging the command to use; so even the initial sync is already done. With this modified code; it should net me the same result…except it will continually update the remote backup rather than just redundantly transferring it. If my WWW directory did something like hit 10 gigs; I’d blow through my transfer in a month.