Preparing the Information Server Enterprise Search node

Before you install IBM® InfoSphere® Information Server, you must prepare the Information Server Enterprise Search node of your cluster.

About this task

Note: This process applies to Version 11.7. In version 11.7.1, a new microservices tier is available. The microservices tier consists of IBM Watson Knowledge Catalog, Enterprise Search, and the Information Governance Catalog and Information Analyzer console. For information about preparing the microservices tier, see Preparing the Information Server 11.7.1 microservices tier node.
Complete the following steps on the Information Server Enterprise Search node of your cluster to prepare it for installation of IBM InfoSphere Information Server with Enterprise Search.

Procedure

  1. Configure the local host lookup for the Information Server Service Tier node if it is not already configured.
    1. To check if you can perform a host lookup for <Information Server Service Tier Node>, run the following command: 
      nslookup <Information Server Service Tier Node>
    2. If the local host lookup is not configured, append the following line to the file /etc/hosts: 
      <IP address> <Information Server Service Tier Node> <Short host name>
  2. Install the following packages if missing.
    • lvm2
    • wget
    • python
    • curl
    1. To check if a package is installed, run the following command: 
      yum list installed <Package Name>
    2. To install a package if it is missing, run the following command: 
      yum install -y <Package Name>
  3. Check devices and partitions.
    • Version 11.7.0.2

      Check the devices to make sure you have a raw, unformatted disk or an unformatted partition. The minimum required space is 100 GB.

      Display information about all the available block devices including an unformatted disk on your system by running the following command:
      lsblk
      See the following example. The system or output might look different depending on your environment.
      [root@sombra-node1 ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
vda 252:0 0 250G 0 disk
├─vda1 252:1 0 1G 0 part /boot
└─vda2 252:2 0 244G 0 part
├─rhel-root 253:0 0 240G 0 lvm /
└─rhel-swap 253:1 0 4G 0 lvm
vdb 252:16 0 200G 0 disk
├─vdb1 252:1 0 250G 0 part <<<<<<Raw unformatted partition <<<<<
vdc 252:32 0 250G 0 disk <<<<< Raw unformatted disk <<<<<

      Confirm that you have a raw, unformatted disk or unformatted partition. This information will be provided as input for the Enterprise Search installation.

      Raw unformatted disk /dev/vdc OR raw unformatted partition /dev/vdb1.

      The installer will take care of formatting.

    • Version 11.7.0.1 or 11.7.0.0
      1. View all partitions in all devices by running the following command:
        parted -l
      2. Check all partitions in a specific device by running the following command:
        parted <device> print all
  4. Configure the Docker volume by using one of the following methods.
    Note: Skip this step if you are using IBM InfoSphere Information Server version 11.7.0.2 or later.
    • Use a spare, unformatted device.
      1. Create a partition table by running the following command:
        parted <Device> mklabel gpt
      2. Create a partition of 100 GB by running the following command:
        parted <Device> mkpart primary 0GB 100GB
    • Reuse an existing mounted disk partition.
      1. Unmount the existing disk partition by running the following command:
        umount <Mount Point>
      2. Comment the corresponding entry for <Mount Point> in the file /etc/fstab, otherwise the machine will fail to reboot.
      3. Reformat the disk partition to remove previous formatting by running the following command:
        pvcreate -y <Partition>
    • Create a 200GB disk partition on <Device> in the space available after the first disk partition of 200GB.
      1. Check to confirm that the planned space is not already allocated by running the following command:
        parted <Device> print
      2. Create the partition by running the following command:
        parted <Device> mkpart primary 200GB 400GB
  5. Ensure that the planned private IP address range for the Weave network, which Kubernetes uses to allocate dynamic IP addresses as needed, does not clash with the planned IP address range for Docker and is available. The default value is 10.32.0.0/12. The planned IP address range for the Weave network cannot be a public IP address range.
    Note: Docker uses 172.17.0.0/16, which cannot be changed.
    1. Check that the network IP addresses do not conflict with Weave and Docker on the Enterprise Search node.
      1. Query the existing private IP address ranges that are in use by running the following command:
        ip route show
        The output should look similar to the following example:
        default via 10.1.1.254 dev eth0 proto static metric 100
10.1.0.0/16 dev eth0 proto kernel scope link src 10.1.1.207 metric 100
      2. Ensure the Weave IP range that was provided as input during the installation is not already in use.
      3. The Enterprise Search installation uses the following IP address as the default for Docker. Make sure you do not use this IP range also.
        172.17.0.0/16 dev docker0 proto kernel scope link src 172.17.0.1
    2. Run the following command to remove the IP address range for the Docker and Weave networks if they are still assigned after the cleanup script, /opt/IBM/UGinstall/cleanUGInstall.sh, was run to remove the existing installation. 
      ip route del <IP Range>
  6. Configure the services cluster IP address range.
    1. Ensure that the services cluster IP address range does not conflict with the existing host network or with the network_cidr.
    2. Ensure that it is a valid IP address range.
    3. Ensure that the IP address range does not start with 172.17 because 172.17 is used for the Docker IP address range.
    4. Ensure that a public IP address range is not used.
  7. Check to see if IP Forwarding is enabled.
    1. Check the file /etc/sysctl.conf, by running the following command: 
      sysctl -a | grep ip_forward
    2. Enable IP Forwarding if net.ipv4.ip_forward = 0
    3. Edit the file /etc/sysctl.conf, and change it to net.ipv4.ip_forward = 1
    4. To load the settings that were made to the file /etc/sysctl.conf, run the following command: 
      sysctl -p

      The configuration setting might also be in any configuration file in the /etc/sysctl.d directory.

  8. Set up a root passwordless SSH connection from the Enterprise Search node to itself.
    1. If the file /root/.ssh/id_rsa.pub does not exist, you must create the SSH Keys by running the following command: 
      ssh-keygen -f /root/.ssh/id_rsa -q -N ""
    2. Copy the SSH public key to the Enterprise Search node by running the following command: 
      ssh-copy-id -i /root/.ssh/id_rsa.pub root@<Enterprise Search Node>
    3. Validate the root passwordless SSH connection to the Enterprise Search node by running the following command: 
      ssh root@<Enterprise Search Node> hostname
      Note: This step is optional for version 11.7.1.
  9. Ensure that port 443 is used only by the HAProxy load balancer, for a full suite installation. The port must be used for this purpose in order for the IBM InfoSphere Information Server Enterprise Search ingress-nginx-controller to function correctly.
  10. Ensure that the swap space is disabled in order for the pods to come up on the IBM InfoSphere Information Server Enterprise Search node after the reboot of the IBM InfoSphere Information Server Enterprise Search node.
    Before disabling the swap space, first check the /etc/fstab file to ensure that the swap space is not mounted again after reboot. For example, the /etc/fstab file might contain an entry similar to the following example.
    /dev/mapper/rhel-root / xfs defaults 0 0 UUID=.... /boot xfs defaults 0 0 /dev/mapper/rhel-swap swap swap defaults 0 0
    To disable the swap space, run the following command:
    swapoff ‐a
    Note: The swap space must be left disabled after installation.
  11. Take into account the following information about the Information Server Enterprise Search installation directory.
    • You can specify any absolute path as the Information Server Enterprise Search installation directory. If the given directory exists already, the installation is done in the provided directory.
    • If you create the installation directory, the directory must be empty at the time of installation.
    • If you do not create the Information Server Enterprise Search installation directory before the installation, the installer creates the specified directory location.
  12. Check the static host name on the Information Server Enterprise Search node by issuing the following command: 
    hostnamectl --static
    For example, static hostname: jsmith-1171-ug-2. If you are using a short name for the host, then set it to the fully qualified domain name (FQDN) by issuing the following command:
    hostnamectl set-hostname `hostname -f`
    The fully qualified domain name will be set as the static host name and will be the input to the installer for Information Server Enterprise Search host name.
  13. Verify that the Host lookup (nslookup) of the local host on the Information Server Enterprise Search node points to the local host and that it is not registered in DNS to any other name by issuing the following command: 
    nslookup localhost
Server: 172.16.200.52
Address: 172.16.200.52#53
    If the host name is registered to any other name, you must fix the issue before you can successfully complete the installation.
  14. Verify that the following ports are available to the Information Server Enterprise Search node: 
    TCP 6443* Kubernetes API Server
TCP 2379-2380 etcd server client API
TCP 10250 Kubelet API
TCP 10251 kube-scheduler
TCP 10248 kubelet
TCP 10249 kube-proxy
TCP 10252 kube-controller-manager
TCP 10255 Read-Only Kubelet API
TCP 30000-32767 NodePort Services
TCP 6783/6784 Weave Services
TCP 9092 Kafka
TCP 2181 zookeeper