Come for the products,
stay for the community

The Atlassian Community can help you and your team get more value out of Atlassian products and practices.

Atlassian Community about banner
4,328,540
Community Members
 
Community Events
168
Community Groups

How to add the nodes to Jira DataCenter Cluster?

Hi

According to https://confluence.atlassian.com/adminjiraserver/set-up-a-jira-data-center-cluster-993929600.html   ,  I hava set up a Jira DataCenter, But I can't understand how to add a node to this cluster. 20210415132756.png

 

1、Is it necessary to mount the shared home directory in the cluster master node?There is no mount at present, but it can start normally.

2、Does the slave node of the cluster need to install JIRA service separately, and then join the cluster?

3、In the slave node of the cluster, do you need to create and edit files “ cluster.properties ” , is it a shared home directory that only needs to be mounted?

4、In the slave node of the cluster, do you need to keep the default installation directory and the default home directory? Do you want to mount the shared home directory to the default home directory?

5、Is it possible to provide a detailed and complete configuration file of JIRA DataCenter? Especially how to join nodes in cluster.

 

4 answers

My Jira DC can be accessed normally.

But how to display node02 in settings --- system --- clustering?

20210427113216.png

Hi, all nodes should be visible on this page (on any node you should see the status of all nodes).

0 votes

Hi,

The is no concept of master/slave nodes in Jira DC. 


1  Is it necessary to mount the shared home directory in the cluster master node?There is no mount at present, but it can start normally.

Every node needs to have access to the shared home directory

2、Does the slave node of the cluster need to install JIRA service separately, and then join the cluster?

Every node in the cluster need to have: jira service running, jira local home with db configuration and cluster properties (each node must have a different ID);

The node joins the cluster when the service starts, connects to the DB and register itself via DB as a new member of the cluster. 

3、In the slave node of the cluster, do you need to create and edit files “ cluster.properties ” , is it a shared home directory that only needs to be mounted?

^^ each node looks exactly the same: jira process, local home with specific cluster.properties and access to shared home

4、In the slave node of the cluster, do you need to keep the default installation directory and the default home directory? Do you want to mount the shared home directory to the default home directory?

The share home is not a replacement for the local home; each nodes needs both; only few things (example: plugins, attachments, ..) are in the shared home; local home is still used for most of disk operations (indexes, logging, cache replication queues)

5、Is it possible to provide a detailed and complete configuration file of JIRA DataCenter? Especially how to join nodes in cluster.


There is not "complete configuration file of JIRA". Each node is a copy of any other node. Usually the only difference between nodes will be the node id defined in the cluster.properties file. Joining/rejoining the cluster happens automatically when a node with a new node.id connects to the (shared) database.

You may want to try setting up a full Jira DC deployment using AWS Cloudformation templates: https://aws.amazon.com/quickstart/architecture/jira/. This would set up Jira nodes connected to a RDS SQL DB behind a load balancer. 

mac

Hi @Maciej Swinarski 

Can I understand it as multiple JIRA services running independently, balancing the load in a certain way?

20210419104118.png

@yong hu  - exactly; note that we recommend configuring the load balance to have sticky sessions - so once a user is assigned to a node, this node is favoured over other nodes for this user; we do this to re-use existing user sessions; 

usually all the nodes perform the same operations on the same set of data; in DC there is a use case where when we want to do a full re-index, one of the node is selected to perform the re-indexing operation (during this time it should not accept user traffic) and when done the index is distributed across all other nodes;

just a note to your diagram - the DB (mysql) component is also shared across all nodes;

mac

Hi @Maciej Swinarski 

I made some configuration based on this document “ https://developer.atlassian.com/server/jira/platform/configuring-a-jira-cluster/  “,

But when I use ” http://ip “ to visit JIRA, I'm always prompted to enter my account and password.

What is the lack of configuration?20210419165517.png20210419165438.png

 

When I access through IP: port, I only see one node under clustering.

20210419165744.png

Hi @Maciej Swinarski 

just a note to your diagram - the DB (mysql) component is also shared across all nodes;  "

Mysql create databases db_node01/db_node02/db_node03 for jira node01/node02/node03? 

no, there is a single DB, all nodes are connected to a single DB;

@Maciej Swinarski 

thank you.

I think I've solved the database.

Like Maciej Swinarski likes this

Error:

20210415140133.png

 

cluster node2 permission, can write/read:

20210415140231.png

 

cluster share home : /datacenter/jiradc

mount on node2 : /datacenter/jiradc

In the cluster, does the master and slave nodes need to have a uniform version of JIRA?

Suggest an answer

Log in or Sign up to answer
TAGS

Atlassian Community Events