Completing Post Upgrade Tasks

After installing the ArcSight Database and upgrading the ArcSight Platform, you need to update the Database Configuration in the CDF Management Portal, and then clean up the NFS directory. For more information, see:

Reconfiguring the Database Host Address
Cleaning Up the NFS Directory

If you have upgraded Intelligence, you must also perform the following post upgrade tasks:

If you have been using custom SQL loader scripts in the previous versions of Intelligence, you need to apply the custom SQL loader scripts.
If you have been using custom data identifiers, then you need to update the logstash-config-pipeline config map. For more information, see Updating the Logstash ConfigMap for Custom Data Identifiers.

Reconfiguring the Database Host Address

As part of the upgrade, you need to update the Database Host address in order for the database to connect with your cluster.

The ArcSight Database is a separate installation and is not included in the platform upgrade. For more information, see Installing the Database

To update the Database Host address:

Log in to the CDF Management Portal.
Navigate to Deployment > Deployments. > > Reconfigure.
Select the Fusion tab, and scroll down to the Database Configuration section.
Update the Database Host field with the IP addresses of all database nodes, beginning with Node 1 and ending with Node N. For example: node1-IP,node2-IP,node3-IP,... and so forth.
Click Save to activate the configuration changes.

Cleaning Up the NFS Directory

After installing the ArcSight Database and upgrading the Platform, you must clean up the NFS directory to remove instances of existing configuration data before you begin a fresh install.

The ArcSight Database is a separate installation and is not included in the platform upgrade. For more information, see Installing the Database.

You must backup necessary configuration data like custom dashboards, searches etc. to avoid data loss on account of the clean up. For more information, see Backing Up and Restoring Configuration Data for Deployed Capabilities.

Remove the "Reporting" directory under NFS:

rm -rf /opt/arcsight-nfs/arcsight-volume/reporting

Remove the "rethinkdb" directory under NFS:

rm -rf /opt/arcsight-nfs/arcsight-volume/investigate/search/rethinkdb

Restart the pods that require rethinkdb:

NSP=$( kubectl get namespaces | grep arcsight | cut -d ' ' -f1 )

kubectl get pods -n $( kubectl get namespaces | grep arcsight | cut -d ' ' -f1 ) --no-headers=true | awk '/fusion/{print $1}'| xargs  kubectl delete -n  $NSP pod

Applying Custom SQL Loader Scripts

Applies if Intelligence has been upgraded.

If you have been using custom SQL loader scripts in any of the previous versions of Intelligence, then, after the upgrade, the analytics pod enters into a CrashLoopBackOff state. To recover from this state and enable the analytics pod to run properly, do the following:

Launch a terminal session and as a root user, log in to the node where NFS is present.
If you have been using custom SQL loader scripts, review and add the necessary modifications to the new SQL loader scripts present in the following directory:
```
cd /<arcsight_nfs_vol_path>/interset/analytics/vertica_loader_sql/0/1.9.1.x
```
Navigate to the following directory:
```
/<arcsight_nfs_vol_path>/interset/analytics/vertica_loader_sql/0/1.<existing_folder_name>
```
where 1.<existing_folder_name> is the folder where you moved your SQL loader scripts prior to the upgrade.
Update the md5 files with the md5 sums corresponding to the modified SQL loader scripts.

Execute the following commands to restart the analytics pod:

export NS=$(kubectl get namespaces |grep arcsight|cut -d ' ' -f1)
kubectl -n $NS scale deployment interset-analytics --replicas=0
kubectl -n $NS scale deployment interset-analytics --replicas=1