January 14, 2011 at 9:28 am
I have an existing sql 2008 r2 cluster standard edition. I need to add reporting services to the cluster. I tried and it got all the way to the install and then failed with error in attachment
January 16, 2011 at 1:59 am
you dont really want to install SSRS on a cluster node as SSRS is not cluster aware. Install it to a separate server and use the cluster for the database store.
If you really want to install on the cluster node launch the installer and select the first option for
"New SQL Server stand-alone installation or add features to an existing installation"
Select reporting services and install a new instance
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
January 17, 2011 at 12:26 pm
thanks that is what we are doing
January 23, 2011 at 6:23 pm
I believe Perry is steering you in the way of choosing to add a new instance, don't "add features" to the existing installation.
The way we handle "clustered" SSRS is the following:
1. On node A, when you go through the installer, choose new and choose a new name for the instance like "SSRSNodeA".
2. After the install completes, go through the configuration for SSRS and create the virtual directories for the report server and report manager, and create your backend databases within the clustered instance of SQL that is already on the cluster.
3. On node B, when you go through the installer, choose new and choose a new name for the instance like "SSRSNodeB".
4. After that install completes, go into the configuration for SSRS, and create the virtual directories. After creating the virtual directories, go to "Scale-Out Deployment" at the bottom of the configuration.
5. Access SSRS Report manager by going to the SQL Server virtual name URL (http://yoursqlservervirtualname/reports).
To summarize, you are essentially installing SSRS independent of the original SQL Server Database Engine instance. You're installing a separate instance on each node of the cluster, and then attaching both instances of SSRS to the same back-end databases (ReportServer and ReportServerTempDB). When users access through the SQL Server virtual name, they are actually connecting to SSRS on either node A or node B, depending on where SQL Server is running at that time. It just so happens that both instances of SSRS are utilizing the same database instance.
Hope that makes sense.
Check out these links for BOL:
http://msdn.microsoft.com/en-us/library/ms156305.aspx -- Using SSRS configuration manager
http://technet.microsoft.com/en-us/library/ms159114.aspx -- Configure a scale out deployment
Best of luck,
Steve
January 23, 2011 at 11:58 pm
S.K. (1/23/2011)
I believe Perry is steering you in the way of choosing to add a new instance, don't "add features" to the existing installation.
correct, that is exactly what I am saying
S.K. (1/23/2011)
go to "Scale-Out Deployment" at the bottom of the configuration.
It should also be noted that scale out deployments are an Enterprise Edition only feature.
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
January 24, 2011 at 6:37 am
Thank you for pointing out that the scale-out deployment is EE only, Perry. I don't think my reply there will be of much help to the initial poster in that case.
You'd likely then need to do what we did for SQL 2005, and have each instance of SSRS point to its own set of DB's. You'd then need to deploy the reports to both instances of SSRS in order to have DB engine failovers keep your SSRS seamless to your end users.
Instance Node A > ReportServerInstanceNodeA & ReportServerInstanceNodeATempDB
Instance Node B > ReportServerInstanceNodeB & ReportServerInstanceNodeBTempDB
When deploying, deploy to both individual instances:
When accessing the reports:
http://sqlservervirtualname/reports
If SQL Server is running on node A, you'd effectively be hitting http://nodea/reports/ which would be pointing at the ReportServerInstanceNodeA & ReportServerInstanceNodeATempDB databases.
If a failover occurred, you'd then be hitting http://nodeb/reports/ which would be pointing at the ReportServerInstanceNodeB & ReportServerInstanceNodeBTempDB databases.
Hope that gives you a workaround that will work in your environment.
Regards,
Steve
January 24, 2011 at 9:11 am
NLB may work here, never tried it myself.
-----------------------------------------------------------------------------------------------------------
"Ya can't make an omelette without breaking just a few eggs" 😉
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply