When trying to scale pods for confluence, it doesn't inherit configurations

Ahmed Attya
Contributor
January 8, 2025

Hello Guys,

I deployed Confluence on Kubernetes Cluster through helm chart of Atlassian. First, I deployed one replica, and I finished setup from web UI and It is ok.


Now when I change replicas to be 2 from values.yaml file or from kubectl command line and tried to access the new pod, it shows me that the database tables already exist and chose me from restart database configurations or continue and overwrite. in both choices it is not the desired behavior because If I continue and overwrite. it will remove everything I have been done in the first pod.

so why this happens and why doesn't the second pod inherit configurations from the first pod??

1 answer

0 votes
Yevhen
Atlassian Team
Atlassian Team members are employees working across the company in a wide variety of roles.
January 8, 2025

Most likely you don't have shared home configured and the second Confluence pod thinks it's a fresh installation.

Ahmed Attya
Contributor
January 8, 2025

How I configure shared-home.
I just set enabled to be true in values.yaml file

should I do anything else?

thanks

Ahmed Attya
Contributor
January 8, 2025

@Yevhen Hello 

can you check the last reply

Yevhen
Atlassian Team
Atlassian Team members are employees working across the company in a wide variety of roles.
January 8, 2025

Shared storage is something you need to take care of in your K8S cluster. Atlassian Helm charts don't configure infrastructure for you. NFS, EFS, Azure files, Google File storage - it all depends on the K8S vendor and cloud you use.

Ahmed Attya
Contributor
January 8, 2025

@Yevhen for shared home I enabled it from values.yaml file and specify storageClass used in our Cluster and It is already created in both pods
I tested it by adding oracle driver in it and when second pod started I found the oracle driver also in the same shared home in the second pod

but problem is same, second pod think it is new deployment and asks me to overwrite existing database again

any other configurations should I do?


thanks

Yevhen
Atlassian Team
Atlassian Team members are employees working across the company in a wide variety of roles.
January 9, 2025

Do you see confluence.cfg.xml in shared-home?

Ahmed Attya
Contributor
January 9, 2025

@Yevhen 
Yes with all details like license, database configurations

Yevhen
Atlassian Team
Atlassian Team members are employees working across the company in a wide variety of roles.
January 9, 2025

Can you share it (just replace license and any sensitive info). Also please share confluence.cfg.xml from local home.

Ahmed Attya
Contributor
January 9, 2025

shared.jpglocal.jpg

Ahmed Attya
Contributor
January 9, 2025

@Yevhen I uploaded both in local and shared 
can you check please

 

Ahmed Attya
Contributor
January 9, 2025

@Yevhen and should I enable clustering for this case?

Yevhen
Atlassian Team
Atlassian Team members are employees working across the company in a wide variety of roles.
January 9, 2025

Of course, you need to enable clustering to be able to scale Confluence. What you need to do is:

* scale down to 1 replica

* set forceConfigUpdate to true

* set clustering.enabled to true

* run helm upgrade

* ensure you have got clustering props in cfg.xml in shared and local home

* scale up

Ahmed Attya
Contributor
January 11, 2025

@Yevhen  Hello


but I tried when setting forceConfigUpdate to true, every time pod is restarted, it asks me to overwrite database. but I will check again to get back to you

Suggest an answer

Log in or Sign up to answer
TAGS
AUG Leaders

Atlassian Community Events