Configuring Impact as a cluster
IBM®
Tivoli®
Netcool®/Impact can be
configured on Red Hat®
OpenShift® Container Platform as
a cluster of nciservers
for redundancy and scaling.
An Netcool/Impactserver
cluster consists of the primary server and one or more secondary servers. In a cloud deployment,
multiple nciervers
can be run in a cluster within a single cloud instance.
nciserver
can be configured as a cluster; the Netcool/Impact GUI
cannot.The primary nciserver
must be fully running before you spin up more
nciservers
.
To configure Netcool/Impact in the cloud as a cluster, use the following steps:
- Scale the Netcool/Impact
statefulsets to
zero.
oc scale sts <release_name>-nciserver --replicas=0 oc scale sts <release_name>-impactgui --replicas=0
-
Edit the
<release_name>-nciserver
statefulset and increase theNCISERVERS_REPLICA_COUNT
value to the number ofnciservers
needed. This increase can be done in the GUI, or on the command line:export IMPACTBACKENDSTS=$(oc get sts | grep nciserver | awk '{print $1}') oc edit sts $IMPACTBACKENDSTS
Then, search for the
NCISERVERS_REPLICA_COUNT
property and set it to the number of Netcool/Impact servers that you require, for example:- env: - name: NCISERVERS_REPLICA_COUNT value: "2"
-
Wait for the pod to restart, or manually restart by deleting the pod:
oc delete pod <release_name>-nciserver-0
-
Increase the pod count of the
nciserver
to match theNCISERVERS_REPLICA_COUNT
value:oc scale sts --replicas="number" <release_name>-nciserver
Note: It can take a few minutes for all the Impact server pods to come up sequentially. -
Edit the
statefulset
of<release_name>-impactgui
and setNCISERVERS_REPLICA_COUNT
to the value used innciserver statefulset
. This increase can be done in the GUI, or on the command line:export IMPACTGUISTS=$(oc get sts | grep impactgui | awk '{print $1}') oc edit sts $IMPACTGUISTS
Then, search for the property
NCISERVERS_REPLICA_COUNT
and set it to the number of Netcool/Impact servers you require, for example:- env: - name: NCISERVERS_REPLICA_COUNT value: "2"
-
Wait for the pod to restart, or manually restart by deleting the pod:
oc delete pod <release_name>-impactgui-0