Jayce
Fully Optimized
- Messages
- 3,056
- Location
- /home/jason
Times have changed a bit since I posted this last year. My current setup involves an application on the Linux clients (I don't really have any Mac or Windows clients on the LAN anymore) known as BackInTime (I'll call it BIT). BIT is set up to work over SSH (keys paired against the server, etc) and synchronizes data to /media/storage/backups/*user* on an hourly basis entirely in the background without irritating the user.
On to the servers...
Main Server
2x3TB in RAID 1 data drives mounted at /media/storage
Backup Server
1x1TB (soon to be 1x3TB) data drive mounted at /media/storage
Backup server's BIOS is set to turn on when power is restored. Working from a timed surge strip, power kicks on at 1:50 AM. At 2:00 AM, the main server rsync's all important data to the backup server At 4:00 AM, the backup server shuts down via cron. Hours later, power to that port is turned off automatically until 1:50 AM lands again. The purpose of this is to eliminate a 2nd box running 24/7 when it's only really needed for a little bit each day to ensure the backup completes. Rough estimates suggest this two hour window allows for up to 250 GB of data to be synchronized over the gigabit LAN. I could probably downsize the 2 hour window considerably, but I figured it doesn't hurt. It's still saving 22 hours of runtime each day.
On Sundays a different script is ran, which is 100% identical except each rsync entry throws a --delete flag. As a result, six days of the week the server is just syncing what changed. On Sundays it deletes anything that is missing from the source, thereby creating a fresh 'mirror' of raw data to the main server.
I have a 1TB external hard drive that I bring home monthly and synchronize my important stuff to, then I take it to work and store it in my desk.
Everything aside from the external hard drive is 100% automated. No complaints.
On to the servers...
Main Server
2x3TB in RAID 1 data drives mounted at /media/storage
Backup Server
1x1TB (soon to be 1x3TB) data drive mounted at /media/storage
Backup server's BIOS is set to turn on when power is restored. Working from a timed surge strip, power kicks on at 1:50 AM. At 2:00 AM, the main server rsync's all important data to the backup server At 4:00 AM, the backup server shuts down via cron. Hours later, power to that port is turned off automatically until 1:50 AM lands again. The purpose of this is to eliminate a 2nd box running 24/7 when it's only really needed for a little bit each day to ensure the backup completes. Rough estimates suggest this two hour window allows for up to 250 GB of data to be synchronized over the gigabit LAN. I could probably downsize the 2 hour window considerably, but I figured it doesn't hurt. It's still saving 22 hours of runtime each day.
On Sundays a different script is ran, which is 100% identical except each rsync entry throws a --delete flag. As a result, six days of the week the server is just syncing what changed. On Sundays it deletes anything that is missing from the source, thereby creating a fresh 'mirror' of raw data to the main server.
I have a 1TB external hard drive that I bring home monthly and synchronize my important stuff to, then I take it to work and store it in my desk.
Everything aside from the external hard drive is 100% automated. No complaints.
