Backup server almost ready for testing!

I subscribe to Alex Lindsay's school of thought that claims data doesn't exist until it exists in two places. Unfortunately that costs money.

We have about 3TB of user data and 4TB of current videos.  (We will be expanding the video library to 10TB soon I hope.)

Before we can implement a full backup solution, I'd like to duplicate our data.  When talking about this volume of data you have two options: do it right or do it cheap.  We're going to do it cheap! (remember, this is phase 1 and our goal is having data in two places)

The correct way would involve a proper backup SAN, tape backup system with off-site storage, and a bit of could based backup just in case.

The cheap way involves any system that lets me buy off-the-shelf SATA drives and keeps track of file versions.

Here's the Phase 1 setup:

• desktop computer
• 2 extra Intel gigabit NICs
• Drobo Pro
• 4 x 2TB SATA drives
• Crashplan Pro Server

Total cost comes to less than $4,000 which isn't bad for 5TB of iSCSI backup storage that does file versioning and deduplication.  Now I just need to invest another $600 for 3TB more of usable backup storage.  (caveat: single iSCSI nic, single RAID controller, single power supply... the backup is not redundant, only the live data on our MD3000i is)

The next step will be to migrate all of our physical servers to virtual servers and use Veeam to do live deduplicated backups to the same drobo pro.  Then I'm sure there's some way we could set up a simple robocopy script to mirror our entire backup system to another set of drives that get pushed to our colocated server rack downtown.

We will also be looking into cloud-based backups for critical databases and files.

Posted via email from oneseventeen's posterous