1
votes

Our Team has a cluster of 3 node marklogic hosts running in AWS. We are not sure if it had built using cloud formation template or AMI. As per documentation, it says if you are using cloud formation template then managed cluster feature will automatically enabled. I had three doubts :-

  1. Is there any way to find if hosts running in marklogic are created using single AMI or using cloud formation template ?

  2. We wanted to check if "managed cluster feature" is enabled in MarkLogic AWS hosts

  3. Let's suppose if the 3 nodes are built using AMI and joined as a cluster, how we can enabled "managed cluster feature". Do we need to raise a new 3 node cluster using cloud formation template ?

MarkLogic version - 10

1

1 Answers

1
votes

The managed cluster feature is usually on by default in AWS if the cluster was built using the Cloud Formation template. The recommended way to disable it is to use an /etc/marklogic.conf file which allows local settings to override defaults. In this case, the file would contain a line turning the managed cluster feature off:

MARKLOGIC_MANAGED_NODE=0

If this does not exist, the other way to determine if you are using the managed cluster features is to check the /var/log/messages, /var/log/mlcmd and /var/tmp/mlcmd.trace files. At MarkLogic startup, the managed cluster feature will print progress in the logs, or it will print a message indicating that the feature is not in use.

Another way to look is from the EC2 Dashboard side to see if the cluster was created with the MarkLogic Cloud Formation Template, which uses the managed cluster feature. To determine if a host or hosts were created using the MarkLogic Cloud Formation Template, you can look at the tags for the instance/s you are investigating. An instance created by the MarkLogic Cloud Formation Template will include tags with Auto Scaling Group name, the Cloud Formation Stack name and the MarkLogic Stack name.

instance tag tab

The managed cluster feature is intended for use with the MarkLogic Cloud Formation Template, and is not supported on custom deployments.