Upgrading Component Pack to the latest version

Before you begin

These upgrade steps are only valid when upgrading from the latest version of Component Pack. To upgrade from a 6.0.0.4 or 6.0.0.5 deployment, follow the steps in Upgrading from 6.0.0.4/6.0.0.5.
Make sure you have pushed all the docker images you require to your registry as per Pushing Docker images to the Docker registry.
If Pod Security Policy control is implemented in your cluster, make sure you have applied the specific policies required for Component Pack applications as described in the topic Applying the Pod Security Policies for Component Pack.

Upgrading bootstrap

There is no need to run the bootstrap helm chart when upgrading as all of the required bootstrapping has already been done during the initial install. If for some reason you need to run the bootstrap again, you will have to purge the previous bootstrap helm release first before you can install again: helm delete bootstrap --purge. By default, the bootstrap will not remove any secrets or certificates if it has detected that there is a service using them. In case of emergency, you can force the bootstrap to regenerate secrets and certificates by using the helm value env.force_regenerate=true when installing bootstrap.

Important: If you are switching from using Solr to Elasticsearch with Orient Me in this upgrade, you must re-install bootstrap using env.force_regenerate=true to regenerate the Elasticsearch certificates. First, determine the <args_from_initial_install> by running helm get values bootstrap. Once the chart is deleted, these values will no longer be available with this command. Run the following command:

helm upgrade bootstrap extractedFolder/microservices_connections/hybridcloud/helmbuilds/ bootstrap-0.1.0-20190919-022026.tgz \ --set <args_from_initial_install>,env.force_regenerate=true

where <args_from_initial_install> is the same helm values that are set on the current installation. If you don't specify these values, the helm upgrade resets values to their default. You can determine the values used during the initial install by running helm get values bootstrap.

If you previously had Metrics and Type-ahead Search configured with Elasticsearch, you must reconfigure the services to use the new certificates by following steps 4-6 in Enabling Elasticsearch Metrics to connect to a Component Pack server.

Upgrading connections-env

To upgrade connections-env, run the following command:

helm upgrade connections-env extractedFolder/microservices_connections/hybridcloud/helmbuilds/connections-env-0.1.40-20190820-113941.tgz
 \
--set <args_from_initial_install>

where <args_from_initial_install> is the same helm values that are set on the current installation. If you do not specify these values, the helm upgrade will reset values to their default. You can determine the values used during the initial install by running helm get values connections-env.

Upgrading infrastructure

To upgrade the infrastructure, run the following command:

helm upgrade infrastructure extractedFolder/microservices_connections/hybridcloud/helmbuilds/infrastructure-0.1.0-20190925-103016.tgz \
--set <args_from_initial_install>

where <args_from_initial_install> is the same helm values that are set on the current installation. If you do not specify these values, the helm upgrade will reset values to their default. You can determine the values used during the initial install by running helm get values infrastructure.

Upgrading orientme

To upgrade orientme, run the following command:

helm upgrade orientme extractedFolder/microservices_connections/hybridcloud/helmbuilds/orientme-0.1.0-20190925-103231.tgz \
--set <args_from_initial_install>

where <args_from_initial_install> is the same helm values that are set on the current installation. If you do not specify these values, the helm upgrade will reset values to their default. You can determine the values used during the initial install by running helm get values infrastructure helm get values orientme.

Upgrading elasticsearch

To upgrade elasticsearch, run the following command:

helm upgrade elasticsearch extractedFolder/microservices_connections/hybridcloud/helmbuilds/elasticsearch-0.1.0-20190925-020030.tgz \
--set <args_from_initial_install>

where <args_from_initial_install> is the same helm values that are set on the current installation. If you do not specify these values, the helm upgrade will reset values to their default. You can determine the values used during the initial install by running helm get values elasticsearch.

Upgrading mw-proxy (Customizer)

To upgrade mw-proxy, run the following command:

helm upgrade mw-proxy extractedFolder/microservices_connections/hybridcloud/helmbuilds/mw-proxy-0.1.0-20190925-020030.tgz \
--set <args_from_initial_install>

where <args_from_initial_install> is the same helm values that are set on the current installation. If you do not specify these values, the helm upgrade will reset values to their default. You can determine the values used during the initial install by running helm get values mw-proxy.

Upgrading sanity

To upgrade sanity, run the following command:

helm upgrade sanity extractedFolder/microservices_connections/hybridcloud/helmbuilds/sanity-0.1.8-20190918-155203.tgz \
--set logLevel=info,\
<args_from_initial_install>

where <args_from_initial_install> is the same helm values that are set on the current installation. If you do not specify these values, the helm upgrade will reset values to their default. You can determine the values used during the initial install by running helm get values sanity.

Upgrading sanity-watcher

To upgrade sanity-watcher, run the following command:

helm upgrade sanity-watcher extractedFolder/microservices_connections/hybridcloud/helmbuilds/sanity-watcher-0.1.0-20190919-022026.tgz
 \
--set <args_from_initial_install>

where <args_from_initial_install> is the same helm values that are set on the current installation. If you do not specify these values, the helm upgrade will reset values to their default. You can determine the values used during the initial install by running helm get values sanity-watcher.

Upgrading elasticstack

To upgrade elasticstack, run the following command:

helm upgrade elasticstack extractedFolder/microservices_connections/hybridcloud/helmbuilds/elasticstack-0.1.0-20190925-103319.tgz
 \
--set <args_from_initial_install>

where <args_from_initial_install> is the same helm values that are set on the current installation. If you do not specify these values, the helm upgrade will reset values to their default. You can determine the values used during the initial install by running helm get values elasticstack.

Installing/Upgrading cnx-ingress

This helm chart is new from the 6.0.0.7 release and thus should not be deployed with 6.0.0.6. cnx-ingress is required for Orient Me, Customizer, and Sanity. If upgrading from 6.0.0.6, follow the installation instructions in this topic Installing Ingress Controller.

If upgrading from 6.0.0.7 onwards, run the following command:

helm upgrade cnx-ingress extractedFolder/microservices_connections/hybridcloud/helmbuilds/cnx-ingress-0.1.0-20190918-222428.tgz 
\ 
--set <args_from_initial_install>

where <args_from_initial_install> is the same helm values that are set on the current installation. If you do not specify these values, the helm upgrade will reset values to their default. You can determine the values used during the initial install by running helm get values cnx-ingress.

Hardening Persistent Volume Permissions

As part of previous release PV set up, it was required to set the/pv-connections directory up with world read/writable permissions. In the 6.0.0.7 release and onwards, you now only need to set permissions so that owner can read, write and execute. For the customizer PV, you need to set permissions so that others can read and execute. Run the following commands on your storage server to make the change:

sudo chmod -R 700 /pv-connections
sudo chmod -R 005 /pv-connections/customizations

Note: As services run, the directory and file permissions will be changed by the container users.

What to do next

With the Ingress Controller now in play, if you have Orient Me deployed, you will have to update the ports in the httpd.conf file of the IBM HTTP Server as per the steps in Configuring the HTTP server for Orient Me. It also means that you can now remove the orient-web-client (30001), itm-services (31100), community-suggestions (32200), appregistry-client (30285) and appregistry-service (32212) ports from your load balancer if you have an HA deployment, and replace them with the http Ingress port (32080).