There are two ways to manage the SVC, with a GUI and though a Command Line Interface (CLI). This page discusses CLI commands, but it is not intended to be an exhaustive list of those commands and parameters, IBM has provided some good manuals and redbooks for that. This is just intended to give a flavour of the commands and how they could be used to define and manage an SVC cluster.
To set up a new cluster, you need to start in the machine hall with the front panel of an SVC node. This panel has 4 arrow keys to navigate around a menu, a select key to confirm a choice, and a two line panel display. You use the left and right buttons to move between commands, and the up and down buttons to move between command options. Once you get to where you want, you press the select button.
A word of warning, when you create a cluster, at some point you will be given a password to be used for remote access, and that password is displayed for just 60 seconds, so make sure you have pen and paper ready. You will also need an IP address for the cluster.
Navigate up and down until you find the Cluster command, then left and right until you find the Create Cluster option, and press Select.
You will then be prompted to enter the IP address, which you change by using the up/down and right/left buttons until you get what want. Note you have to change each of the 4 parts of the IP address separately. When you are happy, press Select again
Now the screen will prompt you for the Gateway IP address, which you change as above, then press Select.
Then you need to navigate right/left until you find the Create Now option, press Select and you create the cluster with a single node.
The top line of the display will then say PASSWORD, and the bottom line will display the password. Write it down, and note that it is case sensitive.
The SVC manages communication with the outside world with SSH Public and Private keys. Before you can start to use either the Controller Console or a CLI interface on your workstation, you must generate the SSH keys, store the public key on the SVC cluster and the private key at the remote end. Check with your documentation for details on how this works.
Once you create the cluster, use the SVC Controller Console and from the front screen, click on Add SAN Volume Controller Cluster. On the next screen, enter the IP address that you used above, click OK and you will be asked for a Userid and password. The Initial Userid is Admin, and the password is the one that you wrote down earlier. Now you can complete your cluster configuration by resetting your password, activating your licenses and adding a service IP address. With the base cluster configuration now complete, you can start to use the CLI interface. The userid for the CLI will be admin, and the password will be as you reset it above.
Using CLI Commands to manage the SVC
There are two basic types of SVC CLI commands:
svcinfo will list out information about the SVC and its components
scvtask will actually make changes to the SVC
Adding nodes to a cluster
Now that you have defined a basic cluster, you will want to add more nodes to it. Start by listing out all the unattached nodes
svcinfo lsnodecandidate
this will list out all SVC nodes that have not been assigned to clusters yet.
The next task is to add one of these nodes to your cluster, and you identify it by either its WWN or panelname, which is straightforward, except that they are called id and node_cover_name in the lsnodecandidate listing.
So you add a new node with either
svctask addnode -panelname node_cover_name -iogrp N
or
svctask addnode -wwnodename id -iogrp N
the N in -iogrp is an integer between 0 and 3. Remember that an IO group consists of two, and only two nodes, so the first node you add will go to iogrp0 and will be paired with the original node. Add extra nodes in pairs.
Managing mDisks and MDGs
An mDisk corresponds to a LUN as presented by the controller of the attached storage subsystem, so you don't create mDisks, you detect them. To do this, use the command
svctask detectmdisk
then follow with
svcinfo lsmdisk -delim|
to see what disks have been picked up. The -delim| parameter separates the fields in the command output with a | rather than a lot of spaces, so the output will hopefully not span lines.
Next you need to define a managed disk group with the command
In this case, I'm creating an MDG called MDGDS4FC001 with a 64 MB stripe size (the default is 16MB) and adding an mdisk called mdisk2 to it. You can add more mdisks by separating them with commas in the command above or you can add them to an existing MDG with the command
svctask addmdisk -mdisk mdisk3 MDGDS4FC001
To check out the status of a managed disk group use the command
svcinfo lsmdiskgrp -delim|
Managing vDisks
Now you have your managed disks placed into a managed disk group and all the space split into 64MB extents. Next you need to format those extents up into a virtual disk that you can present to a host. To do this, use the command.
Here I'm creating a 500GB VDisk called VDAIX000 from the extents in MDGDS4FC001, and defining it to the SVC nodes in IO group 0.
Once you define some VDisks you can list them out with
svcinfo lsvdisk -delim|
or check out a specific VDisk with
svcinfo lsvdisk VDAIX000|
Managing Hosts
Next, you need to define your hosts to the SVC, but before you can do that, you need to know the WWPNs for the HBAs. Exactly how you do that will depend on what the host types are, so I'm going to assume that you know how to do this. There is a SVC Host Attachment Guide that will help you here.
Start by listing out all the unattached HBA ports that are zoned to the SVC.
svcinfo lshbaportcandidate
The port names will be 16 HEX digits like
10000000DE1A34F2
10000000DF1045A2
Check that you can see the WWPNs that you are expecting, then define the host using
svctask mkhost -a name P1201234 -hbawwpn 10000000DE1A34F2
Here I'm creating a host definition for server P1201234 and connecting it with one port. You can add both WWPNs in the command above by separating them with a colon, or you can add the second using