I have two unraid servers running now which I am loving btw; one which has a lot of crap that I don't use very often and the other is my main media server.  Total between both servers is 12tb.  My plan of using windows home server 2011 with drive pool to robocopy/mirror the drives from both servers to the drive pool as a backup copy isn't working.  My next option is to build a third unraid server with enough space for both servers and I want to mirror the drives on both unraid servers to this third server and be able to tell which data is on what server.  Is this possible with two different servers/arrays and if so how would I go about doing it?  I know absolutely nothing about rsync.  I can spell it and I've read what it does but nothing else.

 

Before you say, why don't you just combine both servers into one.  Well, number 1 reason is that the first server only spins up for monthly parity checks or if I want to copy data to it that I won't be using much.  My main server spins up a lot and I don't want it stored with a bunch of crap.  Also that's a lot of information to go wrong should anything happen so storing this type of data separately is in my opinion best for me.

If you plan on having the third server be double the capacity of the other two server's it's entirely possible.

 

 

The easiest way to do this is to build the third server to be the largest.

Then have disk 1 thru x on server 1 rsync to disk 1 thru x on third server

then have disk 1 thru y on server 2 rsync to disk  x+1 thru y on third server.

 

 

For example if you have 12 1tb drives on one server.

you can rsync disk1-12 to third server

and if you have 12 1tb drives on the second server.

you can rsync disk 1-12 starting with disk 13 on the third server.

 

 

It means you will need a disk for disk duplication. On or off site.

While I would not invest in 1tb drives, I probably would not combine two disks to one on the backup server due to the complexity of it all.

 

 

It's possible, I just wouldn't do it.

 

 

If you expand disk1 on server 1 to a larger drive, you will need to do the same on the third backup server.

It's simple, just more costly.

 

 

Where as if you expanded server 1 and server 2 to the max you can rsync across each server

I think I understand what you're saying and yes the third server would be more than double capacity of both server 1 and 2.

 

Say if I have all 750gb drives on server 1, and say a mix of 1.5tb and 2tb drives on server 2, then on server 3 I would need basically the same capacity drives equal or larger to server 1 and 2 to rsync to otherwise I'm wasting space because I can't rsync say two or three 750gb drives on server 1 to one 2tb drive on server 3, etc.? 

 

My plan originally on my home server was to create a folder called server1 and one called server2, then under that folder I would create folders called disk1, disk2, etc. and robocopy data from those drives to those folders as long as the disk I was copying to was large enough.  You're saying I couldn't do that using unraid on server 3 using rsync, correct?  Only way would be basically same or larger disk to disk?

 

I really do appreciate your help!

Some people have had success making shares named for drives when the backup servers drive sizes don't match up to drive sizes for the source server.

 

Make a share on the backup server called something like "S1D1Back" to represent "server 1 disk 1 backup".

Then rsync the disk to a share named for it.

That way, if you need to restore just drive 1, it should be less painful then if you had backed up an entire user share.

 

I am doing it similar to this since my main server is all 3TB drives and my backup server is all 2TB drives.

 

 

...

That way, if you need to restore just drive 1, it should be less painful then if you had backed up an entire user share.

 

That sounds like a small point to note but I absolutely agree that backing up the contents of each disk as opposed to each share would be the way to go, as your most likely need to restore from a backup would be in the event of failed disk/s.

I think I understand what you're saying and yes the third server would be more than double capacity of both server 1 and 2.

 

Say if I have all 750gb drives on server 1, and say a mix of 1.5tb and 2tb drives on server 2, then on server 3 I would need basically the same capacity drives equal or larger to server 1 and 2 to rsync to otherwise I'm wasting space because I can't rsync say two or three 750gb drives on server 1 to one 2tb drive on server 3, etc.? 

 

My plan originally on my home server was to create a folder called server1 and one called server2, then under that folder I would create folders called disk1, disk2, etc. and robocopy data from those drives to those folders as long as the disk I was copying to was large enough.  You're saying I couldn't do that using unraid on server 3 using rsync, correct?  Only way would be basically same or larger disk to disk?

 

I really do appreciate your help!

 

 

it's possible to consolidate a couple 750GB's to 1 2TB drive on the other side.

But then you have to remember how you have it laid out with backups.

 

 

If you upgrade one of those 750's now you have to consider where the extra space will be utilized.

 

 

From my experience, rsync did not work as well over user shares.

 

 

All of the drives related to the user share will need to spin up. Then you will need to set the split levels correctly so data is consolidated where you want it.

 

 

If money is the issue do it that way. But if you want something simple to manage, especially in the heat of the moment, disk to disk rsyncs are easy. The issue then is cost. Spindle to spindle.

 

 

Keep in mind these will not be incremental or date driven backups. It's a one to one backup copy. Unless you do not use the --delete option. Then you will always be "adding" files without removing deletes.

 

 

if you wanted to have some form of dated archive, lets say 30 days, archive, then you can use rsync in a different way and you would want to have spindles on the backup server larger then the source server. I do this with some of my folders.

 

 

i.e. my images and documents folder are rsynced over to special dated directories using the --link-dest command.

(see my rsync_linked_backup shell on my google code page).

 

 

This lets me rsync certain folders in a dated manner whereby the only storage that is used is for changes after the initial copy.

It's a cool mechanism and you may want to consider it for your important document folders.

 

 

For simplicty sake, I would do a rsync disk to disk backup and if I needed for important document/image folders, use dated folders. then do a disk to disk rsync with --linked-dest parameter to a larger destination drive. i.e. 750gb to 2TB keeping 30 days of changes available. 

 

 

I know the OP is talking about 2 > 1 rsync but in the event of a 1 > 1 does it matter what size drives you use? Would you even setup a user share on the backup server?

I know the OP is talking about 2 > 1 rsync but in the event of a 1 > 1 does it matter what size drives you use? Would you even setup a user share on the backup server?

 

 

I would not bother with user shares on the backup server.

 

 

As far as drive size on backup server, I would go with what the sweet spot is. (as long as it is the same or larger in size to the original backup drive).