Multiple Servers Sharing The Same Disks

[Badboy]

Hi Guys, 

I would like to turn my desktop into a backup server that I can switch to if the main server goes down.  If I move my disks to an external disk chaise, could both servers share the same disks without issue? 

[JonathanM]

Unclear what you mean, but if you are asking can you move all disks in a server including the boot USB license stick to a new motherboard and CPU, sure, that's a normal way most people upgrade. Since Unraid runs from RAM and "installs" itself fresh every boot, new hardware is often not a big deal, especially if you aren't passing any hardware through to VM's, and both setups use plain HBA's (no RAID enabled) to access the disks.

 

If you are asking if 2 systems can access the same disks simultaneously, the answer is a resounding NO.

[Badboy]

Sorry, I currently have all my disks in a Thermaltake case with motherboard etc.  I would like to move all my disks to an external enclosure so the second server I build can access them, so if my main server goes down  I just flip over to the other one.  Is this doable? 

[JonathanM]

Your question still isn't completely clear, but my answer hasn't changed. If you move ALL the disks and the boot USB to another system it will work fine given the limitations I listed. If you try to connect both systems to the same disks simultaneously that won't work.

[Frank1940]

@Badboy,  I think I see what you are trying to do.  But I think if such a solution exists, it will only be using a redundant server setup assembled by one of the big suppliers of server farms. (And it will not be cheap!)

 

But what type of failure are you trying to prevent?   If uptime on your Unraid server is the priority, you should have only the most basic NAS setups on the planet.  (No Dockers or VMs, very few plugins.)  And when I look at the Forum most of the failures in accessing the files on the array, they are related to the disk drives themselves.  And many of those are related to write failures to a single drive-- Not an actual drive failure!  If you simply connect the disk drives to a new controller without addressing that failure-to-write issue, you will have a system with the Parity drive(s) being out of sync with the physical drives as the second system will not be aware that the first system had disabled one of the physical drives!

[Badboy]

On 10/22/2023 at 10:36 AM, JonathanM said:

Unclear what you mean, but if you are asking can you move all disks in a server including the boot USB license stick to a new motherboard and CPU, sure, that's a normal way most people upgrade. Since Unraid runs from RAM and "installs" itself fresh every boot, new hardware is often not a big deal, especially if you aren't passing any hardware through to VM's, and both setups use plain HBA's (no RAID enabled) to access the disks.

 

If you are asking if 2 systems can access the same disks simultaneously, the answer is a resounding NO.

That's what I was asking, thank you.

[Badboy]

3 hours ago, Frank1940 said:

@Badboy,  I think I see what you are trying to do.  But I think if such a solution exists, it will only be using a redundant server setup assembled by one of the big suppliers of server farms. (And it will not be cheap!)

 

But what type of failure are you trying to prevent?   If uptime on your Unraid server is the priority, you should have only the most basic NAS setups on the planet.  (No Dockers or VMs, very few plugins.)  And when I look at the Forum most of the failures in accessing the files on the array, they are related to the disk drives themselves.  And many of those are related to write failures to a single drive-- Not an actual drive failure!  If you simply connect the disk drives to a new controller without addressing that failure-to-write issue, you will have a system with the Parity drive(s) being out of sync with the physical drives as the second system will not be aware that the first system had disabled one of the physical drives!

I now know it won't work that way. The only other way is to build a completely separate server and have the data sync as it changes.  I think Space Invader had a video on this,  where the other server was setup to turn off after it had updated the data. If the main server ever goes down, you can use the other one.  I guess the short cut is out of the question. LOL...Thanks for the input.

[Frank1940]

I have done this (set up a second server) but I manually sync the data (using rsync) periodically as the data I am concerned about at this point is a massive collection of media files that I would hate to have to try to find and download.   (In fact, some may not even be available any more...) 

 

The bigger problem would be if you are using the server on a daily basis to store files that are updated on a very frequent basis-- say hourly or even shorter time period. 

 

Understand that a parity protected server is not a complete backup solution.  You need multiple copies of irreplaceable files (with, at least, one copy offsite) to accomplish that!

[Badboy]

Hi,

I think I'm going to have to do it the way you have done it. I have a desktop I can convert, with some newer parts sitting here. The major expense will be the GPU for Plex, in case I ever have to switch over from the main server. I like the Nvidia cards. I have a p2200, works great, good for Fileflow compression, low power consumption.  I'm currently using old disks that I have replaced to back all my media files again. Thought it was better then them just sitting there. Not really old (5 or 6 years), they fit in the same server case. The only bad thing is they are not in a raid, so if one dies the whole backup dies.  I figure if it happens, it happens. Still have the main server drives. Thanks for your input, appreciate it.