Questionnaire for Clustering in the Rational solution for Collaborative Lifecycle Management

Boris Kuschel, Jim Ruehlin Jazz Jumpstart Team
Last updated: June 11, 2012
Build basis: CLM 4.0

At the time of general availability of 4.0, a key file is needed to enable clustering. When you contact IBM support to obtain key file for activating the cluster function you will be asked to provide answered to the following questions. IBM Rational support can then better help you plan the right HA environment for your needs and verify the readiness of your intended cluster environment. It is also a good tool for understanding things that need to be taken into consideration for clustering.

Assessment

The answer to all of the following questions must be YES to take advantage of Clustering in CLM 4.0. If the answer to any of the following is NO, you cannot move to a clustered environment at this time.

1. Is High Availability the only purpose supported by your clustering needs?

2. Will the cluster live on a single site (the cluster will not span physical locations or buildings) with low latency connections?

3. Will the clustered environment run on the AIX or Linux flavors specified in the system requirements?

4. Are all CLM 2012 applications and their JTS upgraded to v4.0?

5. Is the high availability use case supported by the organization? Specifically taking into account:
   

   1. It is only intended to support high availability.

   2. Is not currently for the purpose of performance scalability (performance improvement)

   3. Could result in decreased performance if the appropriate recommendations are not followed

   4. It ill require extra hardware and software for multiple installations.

   5. Requires significantly more installation and administration overhead than a non-clustered environment.

6. Is the topology consistent with one of the standard topologies?

7. Are all of the CLM applications hosted on physical servers if using Linux (no virtual servers)?

8. Are all the middleware components and versions you will use supported by the CLM 2012 applications?

9. Do you have WAS ND version 7.0 or 8.0 with latest fix packs applied ( 7.0.0.23 and 8.0.0.3 at time or writing)?

10. Is there someone in the organization who is comfortable administering WAS ND, and is that person assigned to support the CLM cluster?

11. Do you have an IBM Rational Field Sales or ISSR rep assigned to help you move to a clustering environment?
    

12. Do you have an administrator assigned to work with the Field Sales or ISSR rep who will be responsible for supporting the clustered environment?
    

13. Do you have the appropriate hardware or software to adopt the recommended JVM, WAS, and ORB settings? See the clustering entry in the InfoCenter for the recommended settings.

14. If you will have more than one proxy, will you also have a load balancer?

15. Do all of the planned hosts that will support the cluster nodes recognize each other by their host names?

16. Can you set up a Network Time Protocol (NTP) server on one of the nodes that is used by all the other nodes?

17. Are you able to install products as a root user (Linux) or administrator (Windows)?

18. Do you have at least 3 days to dedicate to setting up and validating the clustered topology, after the necessary hardware has been acquired and set up?

19. Are you installing RRDI clustering and are aware that RRDI cluster has its own requirements and should be installed on a seperate node from CLM?

---

Planning and Preperation

Consider doing the Clustering lab of the CLM 2012 Administration Workshop. The configuration you create will not be like your organization’s cluster, but it will give you a basic understanding of clustering and let you do a “dry run” of clustering setup.

Also make sure to read the detailed library article about clustering.

Use the following questions to gather the information you’ll need to set up clustering in your environment.

1. Which CLM applications will be used in the clustered environment, and how many instances of each application will you have?

2. How many total users will be registered in JTS?

3. How many users will be on each application (CCM/RRC/RQM)?

4. How many concurrent users will there be?

5. What is the breakdown of users? For example, how many users will:

   1. Access/create work items

   2. Deliver to SCM

   3. Initiate builds

   4. Run tests

   5. Create/access requirements

6. What type of licensing will be used (floating/authorized/token)?

7. What is the current High Availability solution?

8. What O.S. (type and version) do you plan to use for the clustered environment?

9. What database (type and version) do you plan to use for the clustered environment (for example DB2 HADR, Oracle RAC)?

10. Do you plan to use a high availability database configuration? If so, which one?

11. Do you have a Websphere eXtreme Scale (WXS) farm? If so, do you plan to use it for the CLM cluster? What is your load balancing and proxy strategy?

    1. Do you plan to have redundant proxies (2 or 3)?

    2. Which proxy do you plan to use, Websphere Web Server Plugin or Websphere Proxy?

    3. What solutions will you use to balance the load (Websphere Edge components, F5, etc)?

12. Which level of system availability do you plan to achieve? For information on system availability levels, see Levels of Websphere system availability in the Websphere Redbook on High Availability.

13. What is the hardware each node runs on? Include processor type and speed, number of cores, memory, disk size, etc.

14. Is each node configured to handle the entire load on its own? We recommend that each node is sized the same as an unclustered CLM instance that would run the entire planned load.

15. What are the characteristics of your network environment, such as the node-to-node and node-to-database speed/latency?

---

Copyright © 2012 IBM Corporation