How to reset an ML instance on Docker+Swarm platform?

How to reset an ML instance on Docker+Swarm platform?

book

Article ID: KB0075246

calendar_today

Updated On:

Products Versions
TIBCO Cloud API Management - Local Edition 5.3.0 and lower

Description

ML5.3.0 or lower instance on a Docker+Swarm platform sometimes runs into issues where TM and/or CM lose bindings when the host(s) restart.
After host restart, TM/CM do not come to ACTIVE state.

This article provides steps to reset the instance.

Environment

Docker using Swarm Cluster.

Resolution

Here are the steps to perform this reset:-
- run "docker container ls" and get the container id for nosql

- login to the nosql container "sudo docker exec -it <container_id> /bin/bash"

- run "cqlsh"  and then "use tmg_registry"

- get a list of tables "describe tables"
--------------
describe tables;

zones           platformapi_nodes  oauth_settings  log_nodes
nosql_settings  cache_nodes        clusters        sql_nodes
nosql_nodes     sql_settings       cm_nodes        platformapi_settings
cm_settings     tm_settings        tm_nodes
cache_settings  log_settings       oauth_nodes
---------------------------
- truncate the below tables:
nosql_nodes, cache_nodes, cm_nodes, tm_nodes, log_nodes, sql_nodes

- exit cqlsh and then exit the container.

- go to the manifest folder(used to create this instance) and run the below script to delete the instance:
delete-tml-cluster.sh

- then create the containers as documented here in the "Creating a Mashery Local Cluster":
https://docs.tibco.com/pub/mash-local/5.3.0/doc/html/GUID-CF586E6B-55AC-4839-8B9D-63771DD79CD2.html
(this is to create the containers in the documented order)

Note:
- as the volumes are not changed, the instance will not lose any existing data(endpoints, certificates etc)
 

Issue/Introduction

This article provides inputs on cleaning up and resetting an existing ML instance.