Learn how to enable geo-redundancy for IBM®
Tivoli®
Netcool®/Impact. A
geo-redundant Netcool/Impact cluster has
the primary server on one Red Hat®
OpenShift® Container Platform cluster and the secondary
server on a different cluster.
Before you begin
Ensure that the
impactadmin
password is identical for
both clusters in the geo-redundant pair. Set the
<release_name>-impact-secret
secret to the same value on both
the primary and backup
clusters:
oc create secret generic <release_name>-impact-secret --from-literal=IMPACT_ADMIN_PASSWORD=netcool
Where
<release_name> is the name that is specified for the release in the
Netcool
Operations Insight® custom
resource (CR) YAML file.
About this task
To configure geo-redundancy for Netcool/Impact, complete
the following steps.
Procedure
-
Install the primary Netcool/Impact server-GUI
server pair. Stop the primary Netcool/Impact server after
the installation is completed to prevent the secondary server from finding the primary server and
starting to replicate, during the installation. To stop the primary server, scale the
<release-name>-nciserver-0
primary Netcool/Impact server pod
down to
zero. oc scale sts/<release-name>-nciserver --replicas=0 -n primary
Where
<release-name> is the name that is specified for the release in the
Netcool
Operations Insight CR YAML
file.
- Install the secondary Netcool/Impact server-GUI
server pair. Stop the secondary Netcool/Impact server after
the installation is completed to prevent the secondary from assuming the role of the primary server.
To stop the secondary server, scale the
<release-name>-nciserver-0
secondary Netcool/Impact server pod
down to zero.
oc scale sts/<release-name>-nciserver --replicas=0 -n secondary
- After you install the IBM Netcool Operations Insight on OpenShift operator on the primary
cluster, edit the CR YAML file, which was used to create the IBM Netcool Operations Insight on OpenShift instance. Add the
following lines to the CR YAML file to define geo-redundancy settings on the primary
cluster:
spec:
helmValuesNOI:
nciserver.impactcore.instancesNamePrefix : <primary_server_name>
nciserver.importNCICACerts.enabled: true
impactgui.importNCICACerts.enabled: true
Similarly, edit the CR YAML file on the
backup cluster:
spec:
helmValuesNOI:
nciserver.impactcore.instancesNamePrefix : <backup_server_name>
nciserver.importNCICACerts.enabled: true
impactgui.importNCICACerts.enabled: true
Where the values for <primary_server_name> and
<backup_server_name> are unique strings, different on each cluster that is
used to identify the Netcool/Impact server, for
example, NCI_P and NCI_B. Do not use an FQDN value for these
names.
- Reapply the CR YAML file:
oc apply -f <CR-file-name>.yaml
Where <CR-file-name> is the name of the CR YAML file.
- Rescale the
<release-name>-nciserver-0
primary and
secondary Netcool/Impact server pods
down to zero.
oc scale sts/<release-name>-nciserver --replicas=0 -n primary
oc scale sts/<release-name>-nciserver --replicas=0 -n secondary
Note: Editing the primary and backup CR YAML files causes the NCI pods to restart, hence it is
necessary to scale the primary and secondary Netcool/Impact server pods
down to zero again.
- Set up Netcool/Impact to
communicate across two Red Hat
OpenShift Container Platform clusters. This communication
is achieved by creating a load balancer service to expose the Netcool/Impact ports and
then adjusting the load balancer to direct external traffic to this service. The Netcool/Impact ports that
must be exposed are as follows:
- Derby port: 1527
- Derby replication port: 4851
- NCI server name server port: 9081
- RMI ports in the range of 3000 - 3100
- Configure geo-redundancy for your primary and secondary Netcool/Impact servers, as
described in Configuring primary and secondary Impact servers.
- Configure geo-redundancy for your primary and secondary Netcool/Impact GUI servers,
as described in Configuring primary and secondary Impact GUI servers.
Results
You enabled geo-redundancy for Netcool/Impact.