Introduction

A Shared Storage Pool (SSP) is a pool of storage area network (SAN) devices that can span across multiple Virtual I/O Server (VIOS) instances. SSPs can be configured with Internet Protocol version 4 (IPv4) or Internet Protocol version 6 (IPv6) networks.

This tutorial provides:

  • Detailed steps to configure an SSP cluster with IPv6 networks.
  • Verification of IPv6 over the SSP cluster network.

Prerequisites

This tutorial includes the step-by-step procedure for configuring SSP over IPv6 networks. Users should have basic knowledge of SSP and IPv6 (static and stateless IP configurations). Setting up an IPv6 SSP cluster requires the following prerequisites

  • IBM® POWER6® system or later
  • Dual VIOS – VIOS Version 2.2.2.0, or later, clusters support Internet Protocol version 6 (IPv6) addresses.

Estimated time

The IPv6 cluster creation might take approximately 30 minutes with a preinstalled VIOS.

Steps

Configure the network interface with IPv6

Configure each node which is planned to be added to the SSP IPv6 cluster with the IPv6 network.

In the following example, note that the interface en6 has only IPv4 IP configured.

 $ lstcpip -interfaces

 Name    Address         Netmask          State     MAC

 en1     -               -                detach    00:00:6c:ae:8b:02
 en2     -               -                detach    00:00:6c:ae:8b:02
 en3     -               -                detach    00:00:6c:ae:8b:02
 et1     -               -                detach    00:00:6c:ae:8b:02
 et2     -               -                detach    00:00:6c:ae:8b:02
 et3     -               -                detach    00:00:6c:ae:8b:02
 en6     10.33.0.234     255.255.0.0      up        6c:ae:8b:02:00:60
 et6     -               -                detach    6c:ae:8b:02:00:60
  1. Configure the IPv6 network over en6.

    $ oem_setup_env
    # autoconf6
    
  2. Check and verify if the en6 interface has been configured with IPv6 IP.

    $ lstcpip -interfaces
    
    Name    Address         Netmask          State     MAC
    
    en1     -               -                detach    00:00:6c:ae:8b:02
    en2     -               -                detach    00:00:6c:ae:8b:02
    en3     -               -                detach    00:00:6c:ae:8b:02
    et1     -               -                detach    00:00:6c:ae:8b:02
    et2     -               -                detach    00:00:6c:ae:8b:02
    et3     -               -                detach    00:00:6c:ae:8b:02
    en6     10.33.0.234     255.255.0.0      up        6c:ae:8b:02:00:60
    en6     fe80::6eae:8bff:fe02:60  64      up        6c:ae:8b:02:00:60
    et6     -               -                detach    6c:ae:8b:02:00:60
    

In the case of stateless autoconfiguration, edit the /etc/rc.tcpip file to retain the network interface changes after every reboot.

n(0) root @ alev5: /home/padmin
# cat /etc/rc.tcpip | grep autoconf
# Start up autoconf6 process
start /usr/sbin/autoconf6 ""              <<< uncomment this line

Resolve the hostnames of the nodes

The hostnames of the nodes must resolve to the IPv6 IP addresses with both forward and reverse lookup. Perform the following two steps:

  1. Update the /etc/netsvc.conf file and add or update the following entry in it.

    # tail -f /etc/netsvc.conf
    hosts=local4,local6,bind4,bind6
    
  2. Add entries in the /etc/hosts file of all that nodes that will be part of cluster.

    # tail -f /etc/hosts
    # Internet Address      Hostname        # Comments
    # 192.9.200.1           net0sample      # ethernet name/address
    # 128.100.0.1           token0sample    # token ring name/address
    # 10.2.0.2              x25sample       # x.25 name/address
    # 2000:1:1:1:209:6bff:feee:2b7f         ipv6sample      # ipv6 name/address
    127.0.0.1               loopback localhost      # loopback (lo0) name/address
    #10.33.0.234    alev5.isst.aus.stglabs.ibm.com
    #9.3.113.86     isstnim01.isst.aus.stglabs.ibm.com
    fe80::200:c9ff:fed1:c43e alev6.isst.aus.stglabs.ibm.com alev6
    fe80::6eae:8bff:fe02:60 alev5.isst.aus.stglabs.ibm.com alev5
    

Now the host should resolve to the IPv6 address. It is good to check both forward and reverse lookup and make sure that they resolve to same hostname.

Repeat the above steps on the other nodes that need to be added to the cluster.

Verify hostname resolution of the nodes

  1. From node 1 (alev5):

    (0) root @ alev5: /home/padmin
    # host alev5
    alev5.isst.aus.stglabs.ibm.com is fe80::6eae:8bff:fe02:60,  Aliases:   alev5
    
    (0) root @ alev5: /home/padmin
    # host alev6
    alev6.isst.aus.stglabs.ibm.com is fe80::200:c9ff:fed1:c43e,  Aliases:   alev6
    
  2. From node 2 (alev6):

    (0) root @ alev6: /home/padmin
    # host alev5
    alev5.isst.aus.stglabs.ibm.com is fe80::6eae:8bff:fe02:60,  Aliases:   alev5
    
    (0) root @ alev6: /home/padmin
    # host alev6
    alev6.isst.aus.stglabs.ibm.com is fe80::200:c9ff:fed1:c43e,  Aliases:   alev6
    

Create and verify the cluster

  1. Create the cluster on one of the nodes using the cluster command.

    $ cluster -create -clustername alev5_v6 -repopvs hdisk95 -sp c_pool -sppvs hdisk96 -hostname alev5
    
  2. Verify that the state of the cluster and the pool are OK using the cluster -status command.

       $ cluster -status
       Cluster Name         State
       alev5_v6             OK
    
       Node Name        MTM           Partition Num  State  Pool State
       alev5            9117-MMD021057C77         5  OK     OK
    
  3. Verify that the SSP cluster is using the IPv6 network for communication using lscluster command. In the following example, you can see that IPv6 is the primary address.

       $ lscluster -c
       Cluster Name: alev5_v6
       Cluster UUID: b5c0b74a-d5ea-11e8-8002-00006cae8b02
       Number of nodes in cluster = 1
       Cluster ID for node alev5.isst.aus.stglabs.ibm.com: 1
       Primary IP address for node alev5.isst.aus.stglabs.ibm.com: fe80::6eae:8bff:fe02:60
       Number of disks in cluster = 2
       Disk = hdisk96 UUID = e7bb52db-7f9d-d3d7-61cf-1fa6d33c5c12 cluster_major = 0 cluster_minor = 1
       Disk = hdisk95 UUID = 7b835ba7-4c6b-6e48-c389-2cb367ec4a94 cluster_major = 0 cluster_minor = 1
       Multicast for site LOCAL: IPv4 228.2.0.96 IPv6 ff05::e402:60
       Communication Mode: unicast
       Local node maximum capabilities: HNAME_CHG, UNICAST, IPV6, SITE
       Effective cluster-wide capabilities: HNAME_CHG, UNICAST, IPV6, SITE
    

Add other IPv6 nodes to the cluster

  1. Add the other nodes to the cluster using the addnode command.

       $ cluster -addnode -hostname alev6
       Partition alev6 has been added to the alev5_v6 cluster.
    
  2. Verify that state of the cluster and pool are OK for all the nodes using the cluster -status command.

       (0) padmin @ alev5: /home/padmin
       $ cluster -status
       Cluster Name         State
       alev5_v6             OK
    
       Node Name        MTM           Partition Num  State  Pool State
       alev5            9117-MMD021057C77         5  OK     OK
       alev6            9117-MMD021057C77         6  OK     OK
    
  3. Verify if IPv6 is being used for network communication over the SSP cluster.

       (0) padmin @ alev5: /home/padmin
       $ lscluster -c
       Cluster Name: alev5_v6
       Cluster UUID: b5c0b74a-d5ea-11e8-8002-00006cae8b02
       Number of nodes in cluster = 2
         Cluster ID for node alev5.isst.aus.stglabs.ibm.com: 1
         Primary IP address for node alev5.isst.aus.stglabs.ibm.com: fe80::6eae:8bff:fe02:60
         Cluster ID for node alev6.isst.aus.stglabs.ibm.com: 2
         Primary IP address for node alev6.isst.aus.stglabs.ibm.com: fe80::200:c9ff:fed1:c43e
       Number of disks in cluster = 2
         Disk = hdisk96 UUID = e7bb52db-7f9d-d3d7-61cf-1fa6d33c5c12 cluster_major = 0 cluster_minor = 1
         Disk = hdisk95 UUID = 7b835ba7-4c6b-6e48-c389-2cb367ec4a94 cluster_major = 0 cluster_minor = 1
       Multicast for site LOCAL: IPv4 228.2.0.96 IPv6 ff05::e402:60
       Communication Mode: unicast
       Local node maximum capabilities: HNAME_CHG, UNICAST, IPV6, SITE
       Effective cluster-wide capabilities: HNAME_CHG, UNICAST, IPV6, SITE
    

The lscluster -c result above, certifies that the primary IP address for cluster communication is the IPv6 address (hexadecimal IP representation).

Summary

Setting up an IPv6 cluster is complex because it involves multiple changes in the network configuration files which users might oversee at certain times. The procedure mentioned in this tutorial describes each step in-detail for users to easily configure the IPv6 SSP cluster. This procedure can also be applied to a cluster with any number of maximum possible nodes.