Skip to content

HOWTO remove channel or tenant

Why do that

An existing tenant will not use the PunchPlatform anymore, or a test tenant has to be removed for clarity purposes. This is also useful to a channel, this reulting of a deletion and a creation.

Reminder : if this is only about channel supervision, the command below dust deactivate the channel.

1
$ punchplatform-channel.sh --unconfigure mytenant/mychannel 

Prerequisites

  • Channel must be in a good state, either in STARTED state or in STOPPED. The operator's pp-conf Git repository must be clean and syned to the central repository.

What to do

Stop any data entry in the canal if needed (e.g. stopping TCP syslog forwarding to PunchPlatform); Check that the log rates for the selected channel/tenant are null and no backlog in the PunchPlatform Grafana view;

Stop the channel (or ensure it is already stopped):

1
2
$ punchplatform-channel.sh --stop mytenant/mychannel
$ punchplatform-channel.sh --status mytenant/mychannel

Unconfigure the channel :

1
$ punchplatform-channel.sh --unconfigure mytenant/mychannel

Commit and push changes, the PunchPlatform Admin should yield the tenant or channel as .

To delete the channel, the directory of configuratio sould be removed as well :

1
$ rm -rf $PUNCHPLATFORM_CONF_DIR/tenants/mytenant/channels/mychannel

If you don't want to risk a wrong channel regenration, the generation manifest can be removed as well.

1
$ rm -rf $PUNCHPLATFORM_CONF_DIR/tenants/mytenant/configuration-for-channels-generation/[...]/mychannel.json

Note

after the number of retention hours specified in the punchplatform.properties file (default : 72), the Elasticsearch, CEPH or the Kafka topics are normally empty. Is it then useless but for clarity purposes to delete any remaining Kafka topic here, please contact the MCO Applicative Team.

In case of a full tenant removal, ou may need to reconfigure some platform-oriented datas:

  • removing a tenant-specific Storm cluster from the punchplatform.properties file; in that case do not forget to re-deploy the platform for chacking;
  • removing configuration from the reverse proxy;
  • Closing tenant-specific flow and destroy VMs...