Just got my first job in IT. Turns out my boss likes OSS. Nice! I was just asked to investigate back up solutions. I really haven't found one that fits yet. I like that backuppc doesn't save dupes of files but setting everything as a share seems iffy to put it gently. rsync is great but all of our desktops run windows (cygwin) and it would save duplicate files.

What back up tools can be used in a mixed environment that don't save 5,000 copies of the same file, just update the diffs like rsync, and are secure?

I'd like to back everything to a NAS device localy an have the other office do the same thing. Then send them a copy of our back up and vise-versa. According to my boss our office produces 300GB of data a week. Kinda find it hard to believe. Said our T1 can't handle the bandwith needed to do it. I think it's more likely that he is backing up 300GB a week of duplicate files, but who knows? I'm the new guy.

Can you define your situation and needs a bit more? How many computers of which kinds, network speeds, size and type of NAS, etc. Also, do you just need to back up data and user files? Or do you need system files? And do you just want to be able to recover individual files or have the ability to recover whole partitions or machines?

Is there any possibility of a budget for either hardware or software? Or are you just being asked to do in the hopes you can magically come up with a solution?

As far as I know at present, the major open source backup programs are not capable of deduplication. BackupPC does it (calls it pooling), but has serious limitations that may or may not matter to you depending on your requirements. Bacula has plans for this feature, but they've been in the plans stage for years. So, unless you want to go commercial, you may have to forget this particular feature, take a step back, and try to assess your situation without any preconceptions.

On the commercial side, Retrospect has done that for years and has a strong presence in Windows. But I don't know what your scale is and whether Retrospect will scale for you.

Give us some details and stats on your network and computers and I'm sure people will come up with more answers.

Sorry abut the lack of specifics. I know that in house we support around 100 destop users. Mostly XP pro but we have some 2K and 98 boxes. For servers we have an AS400, Suse, Avya VIOP, MS active directory domain. These are local and I not really sure how many more servers we have but it's rumored that we will be increasing 35% in the next year and I have no clue about the NAS I think it may be on order but I'm not sure.
This is my first IT job and my boss has me taking care of the desktop users right now. I think asking me to investigate back up solutions is his way of seeing where I will go with it. I don't expect to be the one who makes the decisions. Now to be fair to him he did let me set up a box with OOo since we were out of licenses. It also looks like he's open to GPL solutions. Commercial/Free.

Divide and Conquer.

It can get messy covering that much ground with limited resources. If they start throwing money at you, that could change things. What you want is to avoid becoming a one armed paper hanger. Enough metaphors. ;-)

The AS400 may take care of itself. Depends on how big it is, what they are doing with it, and whether they have dedicated staff for it. For starters, I would put that on hold as well as the Avaya. You can come back to it.

Another thing that will require particular attention is the active directory. Likely, they have an admin for that. Again, you can put that on hold and come back to it.

Backing up all the desktops might be over the top as well as too much unnecessary work. Like you said, a gazillion copies of the same thing over and over. What you want is to make the maintenance manageable as well as the backup. What we do is basically 3 different things. First, in as many cases as possible, we image the machines from a standard image held on a server. Second, we have user data on server shares. And, third, we tell people who want their own unique setup on their desktop that they are on their own. They can still save data to a server share, but we don't have the staff to help them with their unique thing.

Our Mac guy uses radmind to image the Macs in classrooms and labs. He has edited the NetInfo database for the Mac image to create a desktop link to the Samba shares on the server. Those links only connect if someone uses them. So, you don't immediately have hundreds of open connections just because all those Macs are running. He can modify an application and shoot out updates in the middle of the night.

Our PC guy used to use Ghost, but I think he moved away from that. He's been using grub with bootp and tftpboot off our Sun server. Images are there as well. He can interrupt the boot and request an update, or he can push them out at night. User and shared data is on a NAS.

For both of those, bringing up a new system just means entering it into our network database so that it will get an IP address and an image.

With that side of things in hand, all I have to worry about is the servers. The one thing I spent money on was an AIT5 tape library. I wanted to be able to keep full and incremental backups running back 6 weeks as well as archives and off site backups. This fills the bill. I then automated the backup process using Amanda. It typically leaves me with lots of free time to work on all the other things I need to deal with.

That's just one way of doing things. It works well for me, because I don't have to deal with all the desktop systems or the duplication of files across them. It works well for the desktop support staff, because they know I have their images and personal data covered.

Anyway, should be fun digging into all this and deciding where to go with it.

For links and reading, you can check:

http://www.linuxquestions.org/bookmarks/tags/backup

Thanks you have given me some good ideas. Reminding me to divide it up more really helps. I don't like the open shares but could set something up with permissions. Like put HR data in one protected share and accounting in their own and just back up what they put up. Reminding me to divide it up more really helps. Heck this might be a go time to get them to allow more OSS and set up webdav.