Windows Clustering help!

[firehawk_1]

Hi. totally new to clustering and reading document/articles one after the other and getting confused.

im trying to setup a demo VM environment. After various attempts I think I got the cluster working. (configured at least)

1xDC
2xNodes
1xVSAN

WS2008R2 EE + all updates.

using StarWind for the iSCSI. Created 1 harddisk/virtual disk using this software (2GB).

its showing that all the computers are connected to the iSCSI.

purpose of this cluster: to demo what happens when Node1 (or 2) goes down, to get the other node up and running and continue doing its job.

Can anyone please tell me (without linking to more documents!) what I am missing and how to set up IIS clustered?


Many thanks!

 

[sk1939]

If your running a virtualized environment it's trivial since you can just duplicate the virtual machine and set the clone as a failover node. If your trying to duplicate the environment that's harder.

 

[firehawk_1]

sure. well, its all using virtual environment. I am setting up an environment, from scratch to demo failover and clustering.

For a start, I just discovered (now I think im goint 5 steps backwards rather than forwards), that if I do a ping to the cluster IP 192.168.1.18 when Node1 in the cluster is rebooting, it is unreachable.

This is from the DC (not in a cluster) and also from Node2 (cluster node)

where am I going wrong?

 

[sk1939]

It sounds like you have a configuration error since it is not failing over to node 2.

A couple things to note:

There is a problem with IIS shared configuration on Windows 2008 Server due to missing privileges for the Application Host Helper Service. For shared configuration to work, you must follow these steps when you set up IIS shared configuration on Windows 2008 Server.

Open an administrative command prompt.

Run the following commands:

net stop apphostsvc

sc privs apphostsvc SeChangeNotifyPrivilege/SeTcbPrivilege/SeImpersonatePrivilege

net start apphostsvc

That's only if you share configurations via software for IIS.

Check out this article and see if it helps:

http://learn.iis.net/page.aspx/213/network-load-balancing

http://technet.microsoft.com/en-us/library/cc754482.aspx

Remember that fail over and load balancing are different, but not mutually exclusive.

If all else fails you can always ask here:

http://forums.iis.net/

 

[firehawk_1]

Thanks.

ok I think I messed up somewhere during configuration. And due to my lack of knowledge, I dont know what im doing in terms of what is a witness disk and what is a high availibility disk and where the shared configuration should go and what file share we are talking about etc...

right, starting for beginning:

I have created 4 disks on my VSAN
the 2 nodes connect to it via iSCSI initiator and auto configured.

then from this point, I believe I went into disk management and initialised them and made the disks as a simple volume.

now in terms of the clustering here, what do I need to do/configure so there is a failover mechinism here but also in regards for having a high availability access to the disk or share?

this is what is confusing me. I realised I didnt have a high availability storage but did have 1 witness disk (15GB) and 2 other disks which were a FileShare cluster.

this probably doesnt make sense for you, so you know what you are dealing with

I just want to know/get steps on how to setup a witness for storage I guess where the storage is still accessible from any nodes in the cluster. does that make sense?

 

[sk1939]

Thanks.

ok I think I messed up somewhere during configuration. And due to my lack of knowledge, I dont know what im doing in terms of what is a witness disk and what is a high availibility disk and where the shared configuration should go and what file share we are talking about etc...

right, starting for beginning:

I have created 4 disks on my VSAN
the 2 nodes connect to it via iSCSI initiator and auto configured.

then from this point, I believe I went into disk management and initialised them and made the disks as a simple volume.

now in terms of the clustering here, what do I need to do/configure so there is a failover mechinism here but also in regards for having a high availability access to the disk or share?

this is what is confusing me. I realised I didnt have a high availability storage but did have 1 witness disk (15GB) and 2 other disks which were a FileShare cluster.

this probably doesnt make sense for you, so you know what you are dealing with

I just want to know/get steps on how to setup a witness for storage I guess where the storage is still accessible from any nodes in the cluster. does that make sense?

This may help you, although it is targeted toward Exchange rather than IIS. Also note that, in Windows Server 2008 R2, the Cluster Shared Volumes feature included in failover clustering is only supported for use with the Hyper-V server role, for when you have multiple Hyper-V servers setup in failover capacity,

http://technet.microsoft.com/en-us/library/bb629541(v=exchg.80).aspx

 

[firehawk_1]

Thanks.
Because of the amount of time wasted from my end (lack of knowledge but eager to learn but less patience due to time constraints!) and because of the Starwind software issues I found, or maybe it was just a bad configuration, I've decided on not dealing with the IIS just yet but will do later.

I deleted the entire Environment and copied over the backup where I just had the DC, NodeA, NodeB and VSAN and then started from that again.

Managed to get a Clustered File System to have folders/files shared in that. and works in terms of the failover (so I reboot NodeA, NodeB kicks in and vica versa)

Created 3 Virtual Disk images from the Starwind software and then made them online, initialised and created NTFS partitions on Node1 before clustering.

I also created an iSCSI LAN (another network adapter) and give each an IP on the VSAN, Node1 and Node2 and from the iSCSI initiator on Node1 and Node2 connected to the images on VSAN via the iSCSI IP Address of the VSAN.

It then automatically created the Quarrum and the 2 other images as clustered disks.

makes sense?

 

[sk1939]

It does, but it's not a simple setup by any stretch of the word. I also generally don't care much for third party iSCSI target software, preferring to go with a hardware client whenever possible just to eliminate potential sources of problems. I admit this is not necessarily my specialty, I'm a network engineer by trade, not a server administrator.

 

[firehawk_1]

Sure. I understand and appreciate your help thusfar.
I agree though about using the hardware but given that its a test environment and a complete virtual environment - its all that I can do (and I dont have time/money either)

 

[KOOLER]

It's a myth hardware is faster / better. Storage hypevisor running on the same physical machines is faster as all reads go from locally cached DAS and all writes are immediately confirmed over the network. With "hardware" or any stand-alone SAN all reads or writes need to touch the wire before actually being completed.

 

[firehawk_1]

I got the environment up and running, but without using IIS...decided overall it probably wasnt the way to go for a number of reasons and was happy with how much I managed on my own. And today, literally 2 hours ago demo'd it and the end client was happy and the idea was "sold" to them for their solution.

Thanks all!

 

[sk1939]

I got the environment up and running, but without using IIS...decided overall it probably wasnt the way to go for a number of reasons and was happy with how much I managed on my own. And today, literally 2 hours ago demo'd it and the end client was happy and the idea was "sold" to them for their solution.

Thanks all!

No problem, glad it all worked out.