4

Well... I messed up. I didn't do enough research and just checked out the headlines then jump head first into the new Scale-Out File Server that Windows Server 2012 R2 brings to the table. At first an Active-Active file server sounds like the only way to go but after you use it, see the performance hit then do more research (which tells you to basically only use it for VERY large files) you might be like me and want to jump ship!

My question is simple. Currently I have a LUN/DISK dedicated to my Scale-Out File Server. The disk (which is of course currently a Cluster Shared Volume AKA CLV) is about 6TB. I want to know can I convert it back to a normal File Server...

My thought was that I would just remove the Scale-Out File Server role and then move the 6TB disk from a CSV back to an available cluster disk. Then just create a regular File Server role and reuse the 6TB disk I have waiting for me.

My concern is that a CSV or the new Scale-Out File Server might have some funky file system or something special and my data might not be there when I bring the regular File Server back online.

Anyone ever try this before?

5
  • This doesn't answer the question, but it might help you avoid it. Are you just trying to use the share currently as a file server? If you are consider creating a vm and place the vhds on the scale out file server. That is the recommended approach
    – Drifter104
    Aug 12, 2015 at 22:36
  • Well that's a different type of approach! :) Creative and it seems like everything is moving towards VMs ... Look at the print servers, now they don't even have a clustered print server, it's all about high available VM print servers. ;) Thanks for the idea. In a non production environment I will test this out at some point and see how raid arrays play. If they play well this is 100% my new method! Where did you see this idea btw? Was is recommend by MS or something?
    – Arvo Bowen
    Aug 12, 2015 at 22:57
  • @Drifter104 I don't know how much it is "recommended" - it does add considerable complexity and complexity does not go down well with a high availability requirement - I certainly would not recommend it.
    – the-wabbit
    Aug 12, 2015 at 23:01
  • @the-wabbit I don't know if I would say it makes it complex... I would actually say the opposite. The way you do print servers now is the same way and the complexity has been takin out of the mix. Now you just have ONE server to worry about for each service/role. The only thing you would really need to manage in the CLUSTER environment is the Hyper-V VMs. And keeping those happy is a LOT simpler then managing multiple roles/services over a cluster. (Just my two cents working in the VM world)
    – Arvo Bowen
    Aug 12, 2015 at 23:17
  • I've lost count how many ms blogs and articles I've read on the subject. This one for example technet.microsoft.com/en-us/library/hh831349.aspx gives examples of what to use and when. SMB3, storage spaces, storage space direct in server 2016. All simplify and open a world of possibilities that have been the realm of expensive SANs. A sofs running with vhds stored on it gives you instant highly available machine.
    – Drifter104
    Aug 12, 2015 at 23:31

1 Answer 1

2

You certainly will lose and need to re-create your shares, but the data on your CSV should not be affected.

CSVFS is (mostly) a regular (NTFS) file system with a different filesystem driver stack (see http://blogs.msdn.com/b/clustering/archive/2013/12/02/10473247.aspx for further info), removing it from the list of cluster shared volumes and adding to available storage will not affect the files and data stored there.

2
  • I was just about to post it. I just tested in a non-production environment. Everything worked out just as I expected with NO side effects. All data was still accounted for.
    – Arvo Bowen
    Aug 12, 2015 at 23:20
  • On a side note... For others trying to to do what I'm trying to do expect a long wait for the disk to be removed from CSV, the larger the drive the longer the time to take the drive offline (remove from CSV). It took about 10 mins or so to remove a 6TB disk from CSV to available. ;)
    – Arvo Bowen
    Aug 12, 2015 at 23:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .