StarWind Software

Hi,
I have taken the steps to correct the NIC configuration I had so they are all seperate and I can get the drives nicely synchronized.
Great we are making progress !!!
Next step. Please keep in mind I am running server 2012.....
I am having difficulty getting the quorum and my VHD storage drives to see each other from both servers..

Here is the question.

Should these be shared storage, that if a change on either server is made to the drive, that change is visible on the same drive on the other server?
My understanding is that it should, perhaps I am wrong.
The documentation is for server 2008 and I have tried to get everything done as per this document, finding my way around in server 2012, perhaps I have got it wrong??

I have run the iscsi initiator and connected to the 4 available connections and also run MPIO and rebooted the servers. It took a while to find where to initialise the disks (new server manager in 2012), they have been initialised and bought online and assigned drive letters.

I cannot see folders I create on server 1 being there on server 2.

It is just about to the stage where I will reload both servers from scratch and start over as I have tried several configurations of netwroking, which required removel of setup steps I had done to date, I am thinking there maybe something that was not cleared correctly in the server 2012 iSCSI or MPIO that is causing perhaps a mis mapped drive??

I am a bit frustrated, but unwilling to give up. I want to have this sorted so I can continue with the testing.
I do not want to setup the virtual machines until I am sure we have everhting correct.

Please advise any thoughts to correct this problem, or if you recommend a reload and start over.
Thanks
Warren.

You do everything correctly. Initially... Then you do a bad thing: you cannot have multi-access to the same SAN LUN (StarWind or not does not matter here)
if your file system (NTFS in this case) is not cluster aware. And it's not. Please follow this thread:

http://www.starwindsoftware.com/forums/ ... t1392.html

So you can do a couple of things:

1) Use cluster-aware file system. SANbolic MelioFS or DataPlow SFS. Both are not cheap.

2) Configure distributed lock manager. I'm aware of only one for Windows - MetaSAN from Tiger Technology.

3) Stick with classic failover CIFS/SMB or NFS share on top of a HA SAN LUN *or* if you're a Windows Server 2012 user you can use Scale-Out File Servers:

http://www.starwindsoftware.com/configu ... or-smb-nas

http://www.starwindsoftware.com/configu ... or-nfs-nas

http://www.starwindsoftware.com/sw-prov ... le-servers

http://www.starwindsoftware.com/sw-conf ... le-servers

FYI 3) is preferred way for now and we're including so-called "NAS Configurator" to do everything listed in this manuals automatically

Hope this helped

Great stuff thanks.
I will start from scratch I think to remove any possibility there are settings left over from the previous attempts to get this working.
I will keep you up to date with progress.
Cheers
Warren.

Sure. Please keep us updated about your progress. We'll be happy to help.

WarrenW wrote:Great stuff thanks.
I will start from scratch I think to remove any possibility there are settings left over from the previous attempts to get this working.
I will keep you up to date with progress.
Cheers
Warren.

Ok I am winning the war, after several battles (due to lack of knowledge as I am an absolute greeny when it comes to this), I now have a functioning Hyper-V Native SAN with a virtual machine on it.
Now to start to play with live migration and backups etc, break things and repair them by taking macines offline etc.
Lots of fun and I am enjoying every minute of the challenge.

Lets see what I can stuff up now!

Once I have a good feel how it all hangs together it will be time to get the purchase orders out.

Speed looks REALLY good so far.

Cheers

Ok,

Tried a live migration and it seems I have something wrong.
The first page of the attched images came up on the source machine... The second is the error on the destination machine. The only error I seem to have when I validate the cluster is a DNS issue of some kind, is this related to the error I am getting when trying to live migrate?

May I as kyou to reproduce this with the iSCSI SAN instead of Native SAN? You can just install one over the other one on each server one-by-one.

Sorry for the delay in getting back to you. (Holiday in New Zealand)
We are trying to use the direct attached storage in the 2 servers, hence the Native SAN choice.

If we are to use the iSCSI don't we need a third server to act as the SAN? I can do this if required, but it gets really messy with the infrastructure we have here at the moment.
I was intending to run just the 2 servers as the cluster.

Is there a problem with what we are wanting to do with the Native SAN software? or is it a configuration issue with DNS and the domain controller?
What I am trying to do is to get a 2 node cluster running in HA using server 2012 as a platform to run our VM's on.

Thanks for the help so far....
Cannot wait to get it going correctly.

Cheers

Hi All,
We are making progress. Appears there are some DCOM services that are blocked by the windows firewall by default. Enabled the predefined rule and whilst I still cannot do a move between hosts (I think this is my error of understanding how this works), I can have a 2003 server running as a vm and whilst connected via a RDP seesion, turn off the machine that is running the vm and it transferred over to the second node and reconnected the RDP session from my client machine and lost no data (apparently) within 5 retries of reconnection from the client.

So I have an HA cluster working. All I have to do now is to learn how to use it properly, break it a few times and restore the VM's etc before going into production.
This is good news.
I am impressed with the speed.
With a bit more testing and understanding (this is my first Hyper-V cluster) I can be confident we will be purchasing the software.

Thanks heaps for the help.

Have a beer on me!
Cheers
Warren.

First of all I am really glad that this issue get solved!

I will answer one question that wasn`t answered from my stand point

If we are to use the iSCSI don't we need a third server to act as the SAN? I can do this if required, but it gets really messy with the infrastructure we have here at the moment.

To be honest that is the great plus of our solution - it doesn`t require to have something additional for its management or checking health status, two nodes are fully functional.

I hope that your testing will pass well.

Please do not hesitate to contact us should you have any additional questions. It would be our pleasure to further assist you!