In this AppNote I will explain how to install NCS on OES Linux. I also describe how to create a shared storage disk to save data on your Novell Cluster.
- To install NCS on the first cluster node, open YaST, system and Novell Cluster Services (NCS). This will open the installation screen. A window may come up showing that you need to install some other Novell services, like NSS. To the question if you would like to install the software for this, just click Yes. If asked, insert the media. For NSS you have to login once into your eDir Tree. Just enter the admin password to continue.
Once NSS is installed properly, you will be asked if you would like to install the RPM package for NCS, Click Continue.
After you click continue, you will be asked a couple of times to enter the media disks. Just insert the disks asked until the NCS LDAP server configuration screen comes up.
- Enter the password for your TREE admin user.
A screen will come up where you can configure your Cluster parameters. First, click the radio button New Cluster. Enter the Contact of the Cluster objects, the Master ip address of the cluster and select your shared storage disk.
Now a window comes up where you can see the ip address of your current server. Make sure it is correct and that you have selected to Start the Clustering Services now.
After clicking Next, a screen comes up that is saving and loading your settings. After a while this screen disappears. Now you have installed the NCS software on the first cluster node.
Now go to the second Cluster Node and do almost the same.
- On the second Cluster Node, Open YaST, system and click Novell Cluster Services (NCS).
After clicking NCS you will be asked if you would like to install NSS, this is just the same as you had with the first cluster node. Just install it and you will see after a couple of CD changes that the Novell Clustering Services Configuration screen comes up.
In this window, click the radio button Existing Cluster. Enter the Context of the Cluster object you entered with the first cluster node.
In the next screen make sure the ip address is correct. It should be the ip address of the second cluster node. Also, make sure the check box is enabled to Start the Clustering Services now.
Click Next to save and load the settings.
- If everything has gone ok, the Novell Cluster Services is now completely installed and should be working.
Now lets check if we can see if we can manage the Cluster.
In earlier versions of Clustering you could manage the cluster from a couple of different applications like Remote Manager, iManager and ConsoleOne. In OES Linux the main management utility is iManager.
So lets open iManager on the first cluster node by entering: http:\\192.168.1.10\nps in your favorite browser.
You will get a login screen. Enter the user name and password for the admin user. In the Tree field, you can enter the Tree name if you have installed and configured SLP, if you haven't done this, just enter the ip address of a server holding a replica of your tree.
In the Left menu click Cluster. Now, in the right panel, insert the name of the cluster or browse to the cluster object and you will see this page.
You can see that both the OES1 and OES2 servers are in the cluster and that the cluster is running (green dot over oes2 and Master IP address Resource is running).
Now it's time to create something so we can test the cluster.
- Now we will create a volume that is running in the cluster. All of the data on the volume must be accessible when the cluster is migrated from OES1 to OES2 or visa versa.
How to create a Cluster Enabled Volume?
Open a Terminal window as user ROOT.
Enter the command: nssmu
This will open the Novell NSS utility.
In the NSS Utility go to Pools and hit enter.
Hit the Insert button
Now you are able to insert the name of a new Pool. I used the name: data
After hitting Enter, you have to select the device where you will create the Pool. In my case this is: sdb.
This is the shared disk that Cluster Node 2 also sees.
Select it and hit Enter.
Now you have to enter the size of the Pool. I used a Pool Size of: 1GB
Hit Enter again.
The Next screen is very important.
In here you can see a couple of settings that can be done to your clustering.
I leave everything as default, but make sure you entered a ip address for the cluster enabled Pool (192.168.1.26 in my case). Also note the Virtual Server Name, this is important if you are going to use scripts to map to the volume in the Pool.
Click Apply to save the settings and hit Esc to go back to the main menu of the NSS Utility.
Now click the Volumes option, hit Insert an give the volume a name.
Click Enter to save it. You will be asked to encrypt the volumes, answer NO. Now you have to select a Pool to create the Volume in. Select the Pool named DATA and hit Enter.
The Volume named data is now created. Note the mount point in the right panel of the screen: /media/nss/DATA
- Now it's time to manage the new volume.
Go back to your iManager screen and note that there is a resource added to the list: DATA_SERVER
This is the cluster enabled volume we just created. As you can see, it is now running (mounted) on OES1. To check this:
Open a terminal window and change directory to /media/nss/ and see if there is a DATA directory.
And there is one.
Go back to the iManager page again and select the check box in front of the DATA_SERVER resource.
Then Click the Migrate TAB to move the Volume to the other cluster node.
After you clicked Migrate, a second screen comes up. In this screen you can select the cluster node where the volume must be migrated to. In our case this is OES2.
Click OK to move/migrate the volume.
Note that the DATA_SERVER Resource is running on OES2.
Look in the table and see that the location of the DATA_SERVER Resource is OES2.
If you now open a terminal window on OES1 and look in the /media/nss directory you will not see the DATA directory anymore, it is now housed on OES2.
- Managing NCS from a command line or terminal window is sometimes very useful. If you can't do some things in iManager, you can do them from the terminal window.
Now I will explain some useful commands you can use if you ever have a problem with your cluster.
If you want to stop the cluster for a software update, this can't be done with iManager. So, if you want to stop the whole cluster, open a terminal window as ROOT and enter:
There are a couple of scenarios you can have. If you just want to stop the Custer services on one Cluster node enter:
This will leave the node from the cluster.
"/etc/init.d/novell-ncs stop" or "rcnovell-ncs stop"
This will stop all NCS related services on the node the command is given on.
To Start NCS again enter:
"/etc/init.d/novell-ncs start" or "rcnovell-ncs start"
this will join the node in the cluster again.
If you just want to stop the whole cluster, enter:
from any of the cluster nodes.
If you would like to see what nodes are in the cluster, from the command line enter:
This will show you a nice info screen with the cluster name, the node in the cluster, the master node, and the epoch number.
As with all implementations there are several ways you can use Novell Clustering Service. I just showed you that clustering a shared volume is very easy to do. Just keep in mind that you are clustering the POOL and not only the Volume, so if you create more volumes in a POOL they migrate all at the same time. If you don't want to do this, create a Shared Pool for every single volume.
There are a lot of parameters you can configure, but I would say, for the most standard things you can leave them as they are. If you want to know how every parameter/setting reflects on NCS, just login to the Novell Documentation site, where everything is documented.
This AppNote shows you how to setup and run a shared volume on NCS on a OES Linux server.