bosh and health_manager

[wei...@gmail.com]

We tried to use BOSH  to deploy CF health_manager; however, when we completed the deployment, we found health_manager did not start. The log shows the following messages:

  /var/vcap/packages/health_manager_next/health_manager_next/lib/health_manager/varz_common.rb:33:in `release': Path [:total] is not held (ArgumentError)
        from /var/vcap/packages/health_manager_next/health_manager_next/lib/health_manager/varz.rb:68:in `block in release_expected_stats'
        from /var/vcap/packages/health_manager_next/health_manager_next/lib/health_manager/varz.rb:68:in `each'
        from /var/vcap/packages/health_manager_next/health_manager_next/lib/health_manager/varz.rb:68:in `release_expected_stats'
        from /var/vcap/packages/health_manager_next/health_manager_next/lib/health_manager/bulk_based_expected_state_provider.rb:137:in `block in with_credentials'
        from /var/vcap/packages/health_manager_next/health_manager_next/vendor/bundle/ruby/1.9.1/gems/nats-0.4.26/lib/nats/client.rb:369:in `call'
        from /var/vcap/packages/health_manager_next/health_manager_next/vendor/bundle/ruby/1.9.1/gems/nats-0.4.26/lib/nats/client.rb:369:in `block in timeout'
        from /var/vcap/packages/health_manager_next/health_manager_next/vendor/bundle/ruby/1.9.1/gems/eventmachine-0.12.11.cloudfoundry.3/lib/eventmachine.rb:256:in `call
   

Could any body point out what's would be the issues, appcloud-130.yml was used? Thanks a lot.

Best regards,

WL

[Bob Nugmanov]

This has to do with bulk api configuration, see previous log entry in hm  log.

Make sure bulk api endpoint is setup properly: that credentials are received and that cc is servicing that endpoint.

Please note that there are new commits in hm that address some of related the issues. However, with proper configuration your problem should be solved.

Let me know whether this helps.

[Yaroslav Ravlinko]

I have the same issue for appcloud-130/131. 

Can you be more specific about 

This has to do with bulk api configuration, see previous log entry in hm  log.

Make sure bulk api endpoint is setup properly: that credentials are received and that cc is servicing that endpoint.

Please give some example 

Понеділок, 18 березня 2013 р. 22:21:00 UTC+2 користувач Bob Nugmanov написав:

[tlab...@pivotallabs.com]

This has been fixed in the latest https://github.com/cloudfoundry/cf-release/tree/release-candidate branch. The release candidate branch is a stable and tested release branch which are the candidates for appcloud releases so should be okay to use. It will be included in the next appcloud release.

For more information about Health Manager please see https://github.com/cloudfoundry/health_manager. The specific commit where this fix was done is https://github.com/cloudfoundry/health_manager/commit/f86ceb50dc6c3b3a57c5e3371ce52a196ac706e1.

- Tim & Bleicke

[Yaroslav Ravlinko]

Unfortunately is not.

I already used release candidate few times and got the same error. Just in case if I created my own release I have no such issue.

Понеділок, 18 березня 2013 р. 18:39:40 UTC+2 користувач wei...@gmail.com написав:

[Maria Shaldibina]

The backtrace shows the old code. Could you please provide the recent error? 

The error should be fixed with this change:

https://github.com/cloudfoundry/health_manager/commit/79ba672dee6cc7b1c8b07c5322a1404daafd3d7d

Make sure your health manager includes this commit.

best,

Maria