cancel
Showing results for 
Search instead for 
Did you mean: 

Clustering On windows / Without SAN

Former Member
0 Kudos

Hi All,

I am wondering about the possibility of maintaining cluster nodes in a common storage - Not in SAN. Which means, the customer is not having a SAN in their environment for them and indeed they are having a rack servers in place. They would like to implement clustering for the ECC in their environment.

However they are not having enough funds to go for the SAN environment.

So my point is if we add the common disk storage for the two nodes (1 & 2 )in the rack or arrays. Is that possible to implement a clustering envionment in their envionment without SAN in place.

Please let me know your suggestions

Note : I am aware of the clustering concepts through the installation guides and also done a clustering installation 2 times in SQL with Windows as a platform (One with 2008 R/2  and another with 2003 ). So I am comfortable with the terms and the installation procedure. But wondering about the possibility without SAN in place


Regards

Vijay

Accepted Solutions (1)

Accepted Solutions (1)

former_member189725
Active Contributor
0 Kudos

I believe you are trying to use the local disks on the blade servers as a shared storage for data and both cluster nodes can access them . Technically this may not be possible as far as I understand because the way SAN disks are configured for cluster nodes to access them , the local disks on a host may not be able to offer the same functionality and flexibility.

I would rather suggest if such a scenario can be discussed with a Windows cluster administrator who can tell you exactly if such a configuration is possible.

I suggest you install DB and CI on 2 nodes and do not implement the cluster as of now with this kind of non-prevalent configuration setup. Once you have SAN in place , set up the system on the cluster with a system copy . I understand its a risk not having a cluster in place for a production environment , but the above scenario might be more painful to you to maintain if at all it works. Just ensure that you have redundancy as far as data is concerned in the 2 systems CI and DB without the cluster in place.

Regards

Ratnajit

Former Member
0 Kudos

Hi Ratna,


Thanks - That make sense.

Instead of that, we planned to make it use of log shipping for now - (Instead of not using the system and keeping it as idle, log shipping will give some kind of confidence for with respect to hardware failures ).

Regards

Vijay

Former Member
0 Kudos

You must understand that there are different solutions for high availablity and disaster recovery. And each of them protects against different failures. Therefore you need to understand what failure you are trying to protect from. If this is server hardware failure you decision is cluster solution you are written about in your first message. If the target of your protection is data corruption then log shipping is preferred solution.

Regards

Roman

Former Member
0 Kudos

Hi,

Thanks for the comment

Completely agreed Roman and also aware of the two different solutions.

As per the first message, we planned to go for a clustering considering the hardware failure - But the hardest part is we are not ready to invest on SAN at this stage. But we have an hardware in place and hence I was looking for any kind of possibility to go for clustering without SAN.

Then after the analysis we understood, its not best to go to the clustering without SAN (Virtual will do - but we need a license for that - That will be an additional cost ). By considering all such things, we left the plan for now to go with the clustering.

However to make use of the system and not keeping it idle or as an alternative for now, we planned to go for the log shipping.

Hope that clarifies our decision

Regards

Vijay

Answers (0)