getting error from mesos_hash_parser

[systeemkabouter]

I tried upgrading my puppet setup from last year to the current module. I tried running it and when that failed I tried again using the slave example in the README. In both occasions the puppet run fails on mesos_hash_parser :

Info: Retrieving pluginfacts Info: Retrieving plugin Info: Loading facts Error: Could not retrieve catalog from remote server: Error 400 on SERVER: mesos_hash_parser(): Wrong number of args given (3 for 1) at /etc/puppet/environments/production/modules/mesos/manifests/slave.pp:105 on node mesos01-dev.localdomain Warning: Not using cache on failed catalog Error: Could not retrieve catalog; skipping run

The offending line in the module slave code :

mesos_hash_parser($cgroups, 'slave', 'cgroups'),

The $cgroups should just be an empty hash as defined in the slave module itself

Running Ubuntu 14.04, Ruby 1.9.3p484, Puppet 3.8.6

Any hints greatly appreciated

[deric]

That's strange. Which version of this module do you use? The error message is slightly misleading, it should accept either 1 or 3 arguments. How do you install Puppet modules? Do you have appropriate version of puppet-stdlib?

Latest modification to this part of code was made in #48. Automated tests are passing on Ruby 1.9.3, Puppet 3.8.5.

[systeemkabouter]

I manage my modules using r10k :

forge "https://forge.puppetlabs.com"

mod 'deric/mesos', '0.7.0' mod 'deric/zookeeper', '0.5.1' mod 'puppetlabs/apache', '1.6.0' mod 'puppetlabs/apt', '2.1.0' mod 'puppetlabs/concat', '1.2.3' mod 'puppetlabs/firewall', '1.4.0' mod 'puppetlabs/mysql', '3.4.0' mod 'puppetlabs/stdlib', '4.11.0' mod 'richardc/datacat', '0.6.1'

mod 'puppetlabs/powershell', '1.0.6' # chocolatey dependancy mod 'badgerious/windows_env', '2.2.2' # chocolatey dependancy mod 'chocolatey/chocolatey', '1.2.1' mod 'puppetlabs/dism', '1.2.0'

[deric]

That should be ok. What arguments do you pass to mesos::slave? Do you configure somehow cgroups settings?

[systeemkabouter]

My original code was :

  class { '::mesos::slave':
    zookeeper      => $master_zookeepers,
    listen_address => $listen_address,
    attributes     => {
      'env' => 'production',
    },
    resources      => {
      'ports' => '[2000-65535]'
    },

    options        => {
      'containerizers' => 'docker,mesos',
      'hostname'       => $::fqdn,
    },

  }

(this is the part I used in a proof of concept las year)

As that failed, I commented it out and tried the example from the README :

  class{'::mesos::slave':
    zookeeper      => 'zk://192.168.1.1:2181,192.168.1.2:2181,192.168.1.3:2181/mesos',
    listen_address => $::ipaddress,
    attributes     => {
      'env' => 'production',
    },
    resources      => {
      'ports' => '[2000-65535]'
    }
  }

Both give the same error here.

[deric]

No other configuration in Hiera? Could you please test latest version from git master? I'm trying to reproduce this on Ubuntu 14.04.2, but everything works as expected.

[systeemkabouter]

The only messos related items below :

messos::master_zookeepers: 'zk://192.168.33.151:2181/mesos zk://192.168.33.152:2181/mesos' messos::master_iprange: '192.168.33.0/24'

I will try the master branch later on

[systeemkabouter]

I'm baffled now.

Tried the master branch version -> worked ok Removed this version and reinstalled released module using r10 -> worked ok

Same setup, same virtual machines, no reboots in between or such.

Maybe it was a weird apache / passenger caching issue, I'm really not sure. I will see what I can do tonight to reproduce the issue.

[deric]

Yeah, it looks like there was older version on master and new fact on client side. Reopen the ticket if the problem persists.