Installation Issues - AWS EKS JIRA DC

[Lloyd Kim]

I am trying to install Jira in AWS on a EKS cluster.  I have gone through all the pre-requisites and everything looks good, when i go to install Jira using Helm Charts provided and making the changes in the values.yaml.  It seems to install, but I am not able to get to the Jira dashboard page.

I see an couple errors that point me to believe that jira is not installed properly, but not exactly sure how to verify on a eks cluster.

Errors I see 

1. in the jira pod I see 

Readiness probe failed: Get "http://192.168.12.159:80/status": dial tcp 192.168.12.159:80: connect: connection refused

2. In the Load balancer health check for the target group I see unhealthy under the nodes health check. 

I appreciate any help.  Thanks

[Jim Knepley - ReleaseTEAM · Accepted solution]

HI @Lloyd Kim 

"Connection refused" results from the target port not being "open" (as opposed to something like a security group, subnet ACL, or some other network control dropping traffic). I think you've got something misconfigured since it's trying to connect via port 80, which requires root privileges to open. Jira normally listens to port 8080 for that reason.

I would tell Jira that to talk to itself, it needs to use the DNS name of the load balancer, regardless of whether the load balancer is Internet-facing or private.

[Lloyd Kim]

Thanks you @Jim Knepley - ReleaseTEAM , I changed the port from 80 to 8080.  Eventually we will be using encryption across, but for now I am just trying to get a working jira on eks to show that it is a solution we can use instead of ec2.......

The port change is now throwing another error. I changed the values.yaml and changed

from 80 to 8080

 # Port definitions

  #

  ports:

    # -- The port on which the Jira container listens for HTTP traffic

    #

    http: 8080

The error I am getting now is

Readiness probe failed: HTTP probe failed with statuscode: 503

and still getting health check error in my target group.  Any thoughts?

[Jim Knepley - ReleaseTEAM]

That's progress... status code 503 is an HTTP response indicating that the service is unavailable, so we know network traffic is reaching a web service, but that web service is broken.

Look at $JIRA_HOME/logs/catalina.out, which contains messages generated during start-up. I suspect you'll see that Jira can't communicate with the database.

[Lloyd Kim]

Thank you @Jim Knepley - ReleaseTEAM - It seems I just needed to wait a little longer.  It seems after sleeping on it, the 503 errors disappeared and the lb health checks are now healthy.  I ran kubectl logs jira-0 to look at the logs and saw that the database was configured properly.  

I appreciate your help.