February 9, 2012 at 6:48 am
Hi, just looking for opinions on SSRS scale out deployment.
What I have is a cluster where the databases will sit and two separate servers for SSRS scale out deployment.
I need SSRS to be flexible enough to work with different applications and lots of different sets of end users.
Option 1.
Install multiple instances SSRS on both the SSRS servers and configure them in a scale out deployment with a NLB. eg
Option 1.
Install multiple instances SSRS on both the SSRS servers and configure them in a scale out deployment with a NLB. eg
SSRS server1 .--|
. .SSRS-1. . . . . . .|
. .SSRS-2. . . . . . .|---->|----|--> http:\SSRS1
. .SSRS-3. . . . . . .|---->|NLB| -->http:\SSRS2
. .SSRS-4. . . . . . .|---->|NLB| --> http:\SSRS3
. . . . . . . . . . . . . . .|---->|----| --> http:\SSRS4
SSRS Server2 .--|
. .SSRS-1
. .SSRS-2
. .SSRS-3
. .SSRS-4
This way I can have different versions (SP/hotfixes) to each SSRS site and each site specific to the user group.
Option 2.
Have two SSRS instances on a scaleout deployment and control user groups access using folders on the SSRS web site eg
SSRS server1. . . --|
. SSRS-1 . . . . . . . .|---- --- --> http:\SSRS1\group1
. . . . . . . . . . . . . . . .|----NLB -->http:\SSRS1\group2
. . . . . . . . . . . . . . . .|----NLB --> http:\SSRS1\group3
. . . . . . . . . . . . . . . . |------- --> http:\SSRS1\group4
SSRS Server2 . . --|
. SSRS-1
This option restricts me to have all user groups to use the same version of SSRS as well as all downtime for maintenance affecting all users at the same time.
I'm inclined to option 1 as that gives me more flexibility, but the initial design here is for option 2. So just looking for opinions/experience of other DBA's on either solution.
May 9, 2012 at 9:59 am
I have now implemented a 2008 SSRS clustered failover solution. If anyone interested read on.
The solution consists of a standard two node active passive cluster with multiple instances.
In brief;
1.SQL installed on the active server using the “Failover cluster installation” wizard with the following services;
Database Services(virtualserver1\instanceX)
Reporting Services(virtualserver1\instanceX, however as it is not cluster aware installs as server1\instanceX, but
all connections redirects to virtualserver1\instanceX)
2.SQL installed on the passive server using the “Add a node to failover cluster” wizard.
3.Installed SQL Reporting Services on the passive cluster using command below with the same instance name (instanceX);
Setup.exe /SkipRules=StandaloneInstall_HasClusteredOrPreparedInstanceCheck /Action=Install
4.Configured Reporting services as a scale out deployment with databases pointing to virtualserver1\instanceX
Now I can access the report manager from both servers and via the virtual server eg
http://server1/reports_instanceX
http://server2/reports_instanceX
http://virtualserver1/reports_instanceX
I was going to add a load balancer so all users come in either via server1 or if that’s down server2 (high availability rather than performance).
However I can do that without the load balancer if the users use the virtual server address as that is only available on the active node.
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply