Hello all,

I will create 2 OM 2012 servers in one Management Group, OM 2012 DBs are stored on 2 clustered SQL servers with fiber SAN.

One of the 2 SQL servers will be OM 2012 reporting.

Now in my scenario I want to create High Availability for OM 2012 Reporting! so can this design work? if yes what is considerations?

Regards,

SQL Reporting Services is not cluster aware so even if you install it on a cluster, it is only installed on a single node.

http://blogs.technet.com/b/momteam/archive/2008/03/28/operations-manager-reporting-supported-ssrs-configurations.aspx

Roger

• Proposed as answer by Alexis YakovlevMicrosoft community contributor Tuesday, October 16, 2012 8:12 AM
• Marked as answer by Graham DaviesMicrosoft community contributor, Moderator Tuesday, October 16, 2012 9:15 PM

Agree with microsoft hopeless guy but As for reporting with high avalability see System Center Operations Manager 2007 Unleashed book by Cameron Fuller - Chapter 10. Complex Configurations. Redundant Configurations. ...implemented using Network Load Balancing. This may help you.

• Marked as answer by Graham DaviesMicrosoft community contributor, Moderator Tuesday, October 16, 2012 9:15 PM

Hi

Be aware that Rogers link for scale out - option 4 - is not supported but it does work.

NLB is also an option. I have done a walk through here of using Netscaler to load balance the SDK service and configure the reporting server to use this. F5 would be another option.

http://www.systemcentersolutions.co.uk/index.php/scom/entry/load-balance-the-scom-sdk-service

Hopefully these answer your question - if you need further assistance then please feel free to reopen and we'll continue to assist.

Cheers

Graham

Hi all,

@ Graham: thanks buddy for your post, I will deploy 2 OM 2012 Servers with the following components:

• OM1: MS+OM Web Console+OM Operations Console+OM Reporting
• OM2: MS+OM Web Console+OM Operations Console [notice that OM reporting will be installed only on one OM1].

OM1+OM2 will be in one OM Management Group "named MG1", MG1 will use SQL 2008 R2 SP1 Cluster [Cluster includes 2 Nodes: SQL1 and SQL2 with one instance named SqlCluster] to store OM databases and logs.

Now I want to make high availability for the following OM components:

• MS
• Web Console
• Operations Console
• OM Reporting
• Application Diagnostic
• Application Advisor

Now the high availability methods I will use for the above components are as the following:

• For MS: I will add additional Management server
• For Web Console: I will use Microsoft NLB. Is this right?
• For Operations Console: I will use Microsoft NLB. Is this right?
• For Reporting: I will install the OM Reporting component on only OM1 server but at the same time I want to make the reporting service/data high available??!! so can this be achieved? if yes then what will be high available? service or data or both? and what is the method and how?
  
• For Application Diagnostic: I will use Microsoft NLB. Is this right?
• For Application Advisor: I will use Microsoft NLB. Is this right?

Notes:

• SQL Cluster uses Fiber SAN.
• I will not use any Hardware NLB like F5, I will use Microsoft NLB.
• I will install SQL Reporting Services on OM1 for the OM Reporting Server
• BTW: what is your opinion in this: http://blog.coretech.dk/kra/how-to-make-operations-manager-2012-rc-scom-2012-fault-tolerance/
  
  
  

Regards,

• Edited by Anas Jalal Wednesday, October 17, 2012 7:52 PM

Hi,

• For MS: I will add additional Management server Right
• For Web Console: I will use Microsoft NLB. Is this right? Right
• For Operations Console: I will use Microsoft NLB. Is this right? Right
• For Application Diagnostic: I will use Microsoft NLB. Is this right? Right
• For Application Advisor: I will use Microsoft NLB. Is this right? Right

• For Reporting: I will install the OM Reporting component on only OM1 server but at the same time I want to make the reporting service/data high available??!! so can this be achieved? if yes then what will be high available? service or data or both? and what is the method and how?

Some inf from book I advised to you 

1. Reporting components do not need to be highly available (as you do not lose data) 

2. The SQL Server Reporting feature is not cluster-aware and we do not recommend you install it on a cluster, because installing both clustered applications and nonclustered applications on a clustered system is not a recommended best practice

3. Microsoft does not recommend installing the SQL Reporting components on a database server already providing OpsMgr functionality (such as the Operations, Data Warehouse, and ACS databases) - But usually we do that ... common practice....further.... We have not seen issues with these components coexisting as long as the hardware acquired for the database and reporting server is sufficient (it's confirmed.. ok)

4. The first step in creating a redundant reporting configuration is installing a single Reporting server with the Data Warehouse Components on a different server (or cluster). This installation places the report databases local to the system where we installed the OpsMgr reporting components.

To make the configuration redundant, manually replicate the reporting databases between the two systems, and then configure NLB on both reporting servers. Finally, change the OpsMgr configuration to use the NLB address instead of the address of the originally installed reporting server. 

something like this http://rburri.wordpress.com/2009/02/18/scom-2007-web-console-and-reporting-server-scale-out-deployment/ (where two reporting servers (in NLB) and you should manually replicate the reporting databases between the two systems). The book names one server as challenger and another as conqueror.)))

maybe something will help!

• Marked as answer by Anas Jalal Monday, October 22, 2012 4:50 PM

To make the configuration redundant, manually replicate the reporting databases between the two systems, and then configure NLB on both reporting servers. Finally, change the OpsMgr configuration to use the NLB address instead of the address of the originally installed reporting server. 

Hello again,

Thanx Alexis for your valuable post, I appreciate this. Regarding manual replicating databases between the two systems!! I did not get this? can you explain it more? notice that my OM servers are using SQL cluster (2 Nodes).

Regards,

• Edited by Anas Jalal Sunday, October 21, 2012 6:55 AM

Hi, I understand that you use sql cluster which is based on MSFC (previously MSCS). But MSFC and NLB is not supported on the same nodes (http://support.microsoft.com/kb/235305/en-us). 

You should use two other servers where you can deploy SSRS and install SCOM reporting components on them. Then each reporting server will have its own Reporting DB which you should replicate between them or this.. when shared DB is on the cluster (http://msdn.microsoft.com/en-us/library/bb522745.aspx). Then you add these servers in NLB and use virtual name for Reporting. 

Also you may 

This also may help 

http://sqlcat.com/sqlcat/b/technicalnotes/archive/2008/06/05/reporting-services-scale-out-architecture.aspx

http://sqlcat.com/sqlcat/b/technicalnotes/archive/2008/10/21/reporting-services-scale-out-deployment-best-practices.aspx

• Marked as answer by Anas Jalal Monday, October 22, 2012 4:50 PM

@Alexis I will have 4 servers as below

Server	Components	Notes
SQL1	SQL 2008 R2 (without SRSS)	Node1 in SqlCluster1
SQL2	SQL 2008 R2 (without SRSS)	Node2 in SqlCluster1
OM1	OM 2012: MS+Web+Console+R/SQL08:SRS for Reporting	Node1 in NLBCluster1
OM2	OM 2012: MS+Web Console+Operations Console	Node2 in NLBCluster1

 

The SQL Cluster will not be on same NLB nodes, and NLB will not be on SQL cluster nodes, every one of them is created on separated nodes.

I will install OM Reporting om OM1 (before this I will install SQL SRSS on OM1). Now notice I will have only one OM 2012 Reporting Server not 2 so there will be one local reporting database for OM Reporting Server.

Thanks a lot

Regards,

• Edited by Anas Jalal Monday, October 22, 2012 5:01 PM

OK, but correct your inf. "OM 2012: MS+Web Console+Operations Console+Reporting" ...As Reporting in the end))))  

So in this situation your reporting is not high available. As if OM1 is not available The Reporting is not available.

High Availability:

1. OM1 (SRRS + OM 2012 Reporting), OM2 (SRRS + OM 2012 Reporting), local reporting databases (synchronised),

2. OM1 (SRRS + OM 2012 Reporting), OM2 (SRRS + OM 2012 Reporting), reporting database is on the SQLCluster1

and check scom reporting settings point to its NLB name  and Reporting Service fault tolerant (like http://blog.coretech.dk/kra/how-to-make-operations-manager-2012-rc-scom-2012-fault-tolerance/)

• Marked as answer by Anas Jalal Monday, October 22, 2012 4:50 PM

High Availability:

1. OM1 (SRRS + OM 2012 Reporting), OM2 (SRRS + OM 2012 Reporting), local reporting databases (synchronised)

2. OM1 (SRRS + OM 2012 Reporting), OM2 (SRRS + OM 2012 Reporting), reporting database is on the SQLCluster1

Thanks Alexis for your post, it is really cool and valuable :D

I think the above 2 high availability strategies are for making OM Reporting high available? right? but in both cases we have to have 2 OM Reporting? (correct me if I am wrong).

Now suppose my 2 OM servers are using SQL cluster for storing OM's DBs, now which of the 2 strategies I can use in this scenario? one of them or both?

Also suppose I will use the strategy number 2, then I understood from you that I have to do manual replication between the 2 systems? so how? using SQL Replication or something else like DFS-r?!

And finally regarding the link you added http://blog.coretech.dk/kra/how-to-make-operations-manager-2012-rc-scom-2012-fault-tolerance/ , with which of the 2 strategies I should use? both?

Best Regards,

• Edited by Anas Jalal Monday, October 22, 2012 5:02 PM

Hi, Yes both are for high availability, and yes in both cases we have to have 2 SSRS + OM Reporting. In your scenario you can use one of them.

As for manual replication it  concerns more manual reports that are not a part of MP imported ...

assume..we have two SSRS (no metter where you place their DBs) + OM Reporting on them (in NLB).  NLBCluster1, NLBCluster2 and SCOMReport.domain.local (name of nlb cluster)

In SCOM Console you set SCOMReport.domain.local as Reporting Server. 

First assume MS connects to the first node (NLBCluster1). we have some reports and are going to import another MP with reports. In that case only NLBCluster1 possesses new reports. But if NLBCluster1 failed and MS connects to another node (NLBCluster2 )..new reports will be dowloaded to other Reporting Server (its DB will be updated)..it's normal.

But if you manually imported some custom reports on one of the nodes their DBs are not synchronized...and you should synchronize manually.... So as it's not a frequent task you can use utility or idias like that (http://www.sqldbatips.com/showarticle.asp?ID=62). Therefore you can create one interface to import necessary reports to any number of Reporting Servers. (simply solution I Think)

"And finally regarding the link you added http://blog.coretech.dk/kra/how-to-make-operations-manager-2012-rc-scom-2012-fault-tolerance/ , with which of the 2 strategies I should use? both? "

You should use both ..as reporting rely on scom sdk as you can see..so you can link them to NLB cluster of MS as well.

• Marked as answer by Anas Jalal Tuesday, October 23, 2012 4:51 PM

Thanks Alexis for your valuable post, well written, understood, cool and delicious : )

but I still have questions: in your last post NLBCluster1 and NLBCluster2 are nodes in SCOMReport.domain.local? right? if yes suppose that NLBCluster1 and NLBCluster2 both have OM MS, OM Console and OM web console and I want to make both of the operations console and web console as well, then SCOMReport.domain.local is enough for making OM reporting/OM console/OM web high available?!

Regards,

Hi, if on both nodes you have all these components including SSRS with local DBs it should...as if one node is unavailable another works with its own components. But it's a little confused because almost everywhere you will use nlbcluster name. So I would prefer two MS in one NLB (for Data Access Service high availability..SCOM console is connected to the Data Access service you know) (physical or virtual) and two independent virtual servers in another NLB for Web and Reporting. As for OM Consoles you can install them whereever you like and where it's necessary to you...as in fact these Consoles will connect to one of MS Data Access Service which is in NLB). 

• Marked as answer by Anas Jalal Tuesday, October 23, 2012 8:49 PM

Thanks Alexis for your valuable posts, I appreciate this..all is clear now.

Best Regards,

https://technet.microsoft.com/en-us/library/hh298606.aspx

Network Load Balancer is not supported for the Operations Manager web console or Reporting server.