Support Questions

Find answers, ask questions, and share your expertise
Announcements
Celebrating as our community reaches 100,000 members! Thank you!

Cannot grow cluster (javax.ws.rs.NotAuthorizedException)

avatar
Expert Contributor

Verified that no changes were made to the director password, related AWS keys, and I have not hit my limit of instances.

 

[2015-04-08 15:44:46] WARN [qtp949320138-46] - o.e.jetty.servlet.ServletHandler: /api/v1/environments/development/deployments/development/clusters/development-cluster
javax.ws.rs.NotAuthorizedException: null

 

Version: Cloudera Director 1.0.2

1 ACCEPTED SOLUTION

avatar
Expert Contributor

I'm happy to hear that you've had good success with Cloudera Director. We appreciate you taking the time to try it out. We take all feedback to heart and strive to improve the product.

 

Have you changed the user/password on your Cloudera Manager instance for this cluster? It looks like the failure occurs during cluster template validation. During this time, Cloudera Director calls out to Cloudera Manager to validate that the product versions match.

View solution in original post

13 REPLIES 13

avatar
Expert Contributor

**bleep**! Yes, changing the manager password to the default admin/admin did resolved the issue. I feel a little stupid but there was no obvious indication that this was the cause, from the errors in the logs and one other thread on this exception I assumed that the API authentication error was coming from the director and not the manager.

 

Can I make two suggestions?

 

1. Please improve the error messages in director. & If users are required to keep the default passwords on their directors and/or clusters in order to manage them this should be made absolutely obvious to the user. The `unknown error` provided to the user is not helpful at all.

 

2. Users should be allowed to change the manager admin password without breaking cluster management in director. Why not allow me to choose an administrator password in director prior to creating the cluster? Or just ask me for the admin username and password when I want to manipulate a cluster. At least that way I can use a secure password and handle user management in the manager separately.

 

I really appreciate your help in resolving this and I hope you take my suggestions to those who might be able to make these improvements. Hopefully the next update to director will be able to manage these clusters.

avatar
Expert Contributor

Yay! I'm glad that we found your issue.

 

1. Yes, the error message was not informative. I'll open a ticket to fix this.

 

2. From v1.1.0, Cloudera Director supports username/password updates for Deployments (i.e., for your Cloudera Manager). You can do this through either curl or the API Console to the /api/v2/environments/{environment}/deployments/{deployment} endpoint. Since only username and password updates are supported through that endpoint, I suggest GETting the deployment JSON, updating the username and password, then PUTting it back to ensure all the other fields match.

 

 

avatar
Expert Contributor

If you could provide me with the ticket URL so that I can track the progress that would be much appreciated. Also if there is already a ticket open about the parcel issue related to Director 1.1.1 and clusters created with 1.0.x and you could provide a link to that ticket as well that would be fantastic.

Thanks again!

avatar
Expert Contributor

Sorry, the ticket is in our internal JIRA. 

We're targeting 1.1.2 and 1.5.0 releases for these fixes.

 

Here's a link to the other message thread discussing the proxy issue

http://community.cloudera.com/t5/Cloudera-Director-Cloud-based/cloudera-director-bootstrap-failed-on...