I'm the system administrator of a small law office (6-8 employees) which stores a large amount of important legal documents (Primarily PDFs, Excel Spreadsheets, and Word Documents) on a simple Windows XP computer, which we call our server.
Each user has full delete privileges (a necessity to be able to work.)
What is a good backup scheme for this sort of system? I am going to run back-up every night, but the difficult question is what to do with moved and/or deleted files. Do we delete them out of the back up? Do we update our back-up to reflect the deletion, but hang onto the deleted file just in case we need it in the future?
Does anyone have an elegant solution to this problem? The data being backed up is about 70GB.
Ideally I'd like a snapshot of the server for each day. As the data is 70GB, this is impossible to do in a direct fashion. Something that keeps a "diff log" from day to day and retains all data, but lets me view the server exactly as it appeared on day X would be perfect.
Edit: Well, I've done some more research, and it sounds like what I'm looking for is fairly common. It's called either "Incremental" or "Differential" back-up schemes, and it looks like the best scheme is one that integrates both.
Does anyone have a recommendation for a piece of software that will help me do this and can be scripted/configured in a fairly flexible way?
In response to "9 Questions you must ask yourself", which was linked to below:
1)How important or sensitive is the data on your systems?
Very important, moderately sensitive.
2) What type of information does the data contain?
Text documents and Scans of Text Documents. Excel Spreadsheets.
3) How often does the data change?
A small amount of the data changes daily. The rest rarely changes. As new Legal Files are opened, the data in those files change until the file is closed, then it might never change again, or just have a little added to it here and there.
4) Can you supplement backups with shadow copies?
Sure. Whatever is needed I can do.
5) How quickly do you need to recover the data?
Speed of recovery isn't an issue. A day would be fine.
6) Do you have the equipment to perform backups?
Right now we only have other office PCs. If something is absolutely necessary, it can likely be purchased.
7) Who will be responsible for the backup and recovery plan?
Me.
8) What’s the best time to schedule backups?
After hours 8pm - 3am
9) Do you need to store backups off-site?
Eventually. Having a good backup scheme in-house will solve 99% of problems.
Start with 9 Question You Must Ask Yourself When Planning a Backup Strategy
A great program is ShadowProtect. this does EXACTLY what you want, you can mount the backup files to get a perfect snapshot of how it was at that time allowing individual file restoration. You can even make a virtual machine from the backed up image in the event of a failure.
You will likely want to use incremental.
Incremental=changes since last incremental. If one file in the chain is lost, the entire chain is lost.
Differential= changes since last full backup. File sizes grow over time but dependency is ONLY on the full back up.
Something else you may want to consider: Some IT consulting companies do BDR(backup disaster recovery) plans which include offsite backups and maintain your backups for a monthly fee normally including a leased server and offsite storage. Trust me, keeping a good backup can be a lot of work.
If you like the idea of something free and open source, BackupPC might be a good option. It's got a quick learning curve, a shiny web interface, and can be placed in a VM (it runs on Linux) if you like.
Installation process: http://www.howtoforge.com/linux_backuppc
Integrating with Volume Shadow Copy (for those pesky locked files -- wouldn't recommend at first): http://www.goodjobsucking.com/?p=62
Example virtual appliance (I can't vouch for its integrity, but would make a good example): http://gotitsolutions.org/files
I use rsync to make snapshot backups, both on-site and off-site. See http://www.mikerubel.org/computers/rsync_snapshots/ for details (that page hasn't been updated in a while, but it still relevant). There are several rsync builds available for Windows. The data sharing between snapshots in this arrangement is at the granularity of files so is not going to be space efficient if you have large files that get regular small updates, but otherwise it makes for a good efficient (both in terms of space and bandwidth used) system.
We use second copy application for same kind of purpose. It first does full backup. Then all inceremental, also it backups the deleted files to another folder in same remote backup location. So you can even bring back the deleted files while having 1 to 1 backups.
As calculation wise. Once it runs with full backup the incremental calculation file size and change based check takes not more then 6 minutes on a vm machine. Its really fast.