Active/Active 2-node cluster with Replication | SQL Server Performance Forums

SQL Server Performance Forum – Threads Archive

Active/Active 2-node cluster with Replication

Can I replicate the databases stored on an active/active cluster shared disk from node1 to node2 and vice versa? Any link that leads to same idea will be great. Any performance degradation if I will be using each node as both publisher and distributor? Thanks,
Jon M
Hi Jon, yes it is possible. the two instances are seperate just as if they were totally different servers. The question is really what are you trying to achieve, as there may be better ways? performance degradation will not be that significant, although it might add up when you’re in a DR situation and therefore running the whole lot on one node. The main danger of active/active is underspec’d servers to run the whole thing on a single node Cheers

Thanks Twan, I will be using replication for my reporting server.<br /><br />Another question, I have disk E for data and disk F for log. Is below setup possible?<br />Folder E:sqldata1 – will be used by sql virtual server 1 <br />Folder E:sqldata2 – will be used by sql virtual server 2<br />Folder F:log1 – will be used by sql virtual server 1<br />Folder F:log2 – will be used by sql virtual server 2<br />In other words, can I use the same physical set of disks (E & F) for both instances and just create subfolders?<br /><br />Can I have only one physical disk for MSDTC and another physical disk for Journal/Quorom for active/active?<br /><br />Do I need to have a separate physical disk for my replication? Or creating a folder in my data disk (E<img src=’/community/emoticons/emotion-1.gif’ alt=’:)‘ /> is fine?<br /><br />Thanks again,<br />Jon M<br /><br />
Jon, I wouldn’t recommend the disk conf. you have listed. You want the SQL instances to be in different cluster resource groups so they can failover independantly of one another. A clustered disk resource must belong to only one resource group, so you cannot have the 2 instances using the same drive for logs/data. You should probably assign E to the first SQL instance and F to the second.
Hi Jon, as mentioned by biged, in an active/active set up with two instances you can’t have both instances using the same disk resources. you could use a separate single disk for quorum and msdtc, BUT and its a huge BUT, the quorum IS your cluster, lose that and you lose the lot. So the Quorum must be raid1, I run msdtc on the quorum disk and in the cluster group. It is fine to do this as long as you don’t have huge volumes of distributed transactions creating a folder in your database data drive is fine, in fact that is the default. the REPLDATA directory may need to be shared out so that the other node can get at it. Use the virtual server name of the instance to refer to the share, so that it fails over properly. Cheers
I have confegured 2 node clustering and a disk is assigned for both the nodes as an instance, but i can able to mount the the bisk on only one node but on 2nd node i can’t. plz give some solution

Welcome the forums.
Wow dusting off 6 years old thread, I see that you are truly following the threads to see if relevant problem is reported.
Anyways coming back to your Q, what error do you see when you try to mount to the second node.
Also confirm the SQL SErver version & operating system version, type of cluster you are using here.

I am trying to add a virtual instance to an existing sql 2000 cluster. When i try to run the install, two things are happening: i am specifying installation of SQL04, but the installation is creating IP and Network Name resources under SQL03. It reports an error creating cluster resources because the group is not in the proper state (sorry dont recall the exact error… will try to post that on monday). Any help will be greatly appreciated.