MS Sql Server installation testing

Installation testing is perhaps one of the most time taking testing category I have ever come across in my professional career. Installation is something that keeps happening for any time range between 20 minutes to 40 minutes, and during this time frame only thing you can do is bird watching. And if you donot have  enough birds in your locality passing time becomes tough.
However , a workaround does exist in this case, during the course of installation you can proceed with some negative testing activities. Negative testing approach can be termed in the sense of parameter validations performed by passing the wrong parameters, such as invalid port, invalid db instance name, invalid collation etc.
Types Of installations :
There are numerous database installation types that can be tested, the major ones are underlined as under:

1. Standalone installation :
  This installation type refers to the one which we generally do during the MSSql Server installation on our personal machines. It however in enterprise installation needs to be assessed for its validation in terms of the port range within which  the installation has to be able to commenced. Other basic features such as the database instance names and all will definitely go fine as that needs to be validated from the sql server's set up file directly. However , we need to have utmost care in the administrator group account we should be able to configure for the database instance getting created, apart from that the service account that should be utilized for running the database instance also needs to be validated. All these validations can be directly performed from the configuration manager, where in the details of the database instance installed is available.Several other validations that are of importance are to verify if the security settings are as per the expectations, such as the error logs have to be captured for what type of events such failed logins only or both successfull and failed logins both. Do not forget to validate the connectivity of the database instance using the port number used at the time of installation.

2. Cluster Installation
 This installation type has some complicated aspects associated with it . There is some restriction in terms of the IP range that can be used for this installation. Generally the best approach to target this installation is the Static IP utilization to consume the service. Get a static IP configured from the lab team on the domain you are performing testing. for my case my case I used some six virtual machines to perform my installation activities. The way I have approached is have three machines blocked just for testing the Cluster installation and the add node to cluster installation. Just have the Windows cluster failover manager installed one any one of the three machine. Do remember to have the DTC running on the machine. Again from validation perspective assure that the properties of the installed database instance is the same as expected with the service being consumed by the account it is expected to. For cluster installation you need to have some cluster network created which you can always do with perfect ease, and pass this as a parameter during the course of installation of the Cluster in the machine.

3. Add Node to a cluster installation :
 Adding node to a cluster is nothing but creating a MSSql server database instance of same name as is already existing within the windows cluster failover manager. So I just consumed the database instance that was created as part of the Cluster installation. So what I meant to say is when you do the add node a pre-requisite is you should have a Cluster installation performed beforehand, first to meet the optimized approach to testing and second to have the test bed created for the add node functionality testing to proceed with. What are the various things we nee to have at the back of mind, use the Cluster network name as used during cluster installation, database instance will be the same as the one used during installation of the Cluster. And this add node has to be done on the same network wherein the cluster network has been configured , that is the machine must be a part of the same network , as already mentioned I hace three machines on one machine I have the Cluster installation done, On remaining two machines I would be performing the add node installation. Once I have performed installation od Add node on second machine its time for the validation which is nothing but go to the main cluster machine where in the windows cluster failover manager has been installed and perform the move instance service to a different node on the cluster. So in my case I would move the service from node one to node two, because on node one I have my cluster installed and node has been added onto the node 2, so I now move the service to node 2 so that the node 2 database instance running and node 1 database instance getting stopped. Do verify that before doing this move node activity and just after having completed the Add node installation , the database service on the added node is installed in Stopped  state. Hoever when we move the service from node 1 to node 2, it is the database service on node 1 that gets stopped and the database service on node 2 that gets started. Similarly add node onto the third node as well ad keep ,moving nodes from one to another among the three machines  namely 1, 2 3 cluster machiness.

4. Enrolment of a Standalone instance to a CMS/MDW server :
5. Security groups enrolment :
6. Access provisioning :
7. Memory range provided to the database instance installed

 To Be Contd.


Post a Comment