error on installation of agent on

[dec-pupop]

When I have the following setup :

vms.yaml :

vms:
- name: master
  box: centos-64-x64-vbox4210-nocm
  roles:
  - pe-puppet-master
- name: first
  box: centos-64-x64-vbox4210-nocm
  roles:
  - pe-puppet-agent

pe_build.yaml : 

---
pe_build:
  version: 3.3.0

roles.yaml :

---
roles:
  pe-puppet-master:
    private_networks:
      - {ip: '0.0.0.0', auto_network: true}
    provider:
      type: virtualbox
      customize:
        - [modifyvm, !ruby/sym id, '--memory', 4024]
    provisioners:
      - {type: hosts}
      - {type: pe_bootstrap, role: !ruby/sym master}

  pe-puppet-agent:
    private_networks:
      - {ip: '0.0.0.0', auto_network: true}
    provider:
      type: virtualbox
    provisioners:
      - {type: hosts}
      - {type: pe_bootstrap}

and using vagrant version Vagrant 1.7.2 with virtual box 4.3.24 I get the following error :

Stderr from the command:

!! ERROR: Puppet Master at 'master:8140' could not be reached.
   Aborting installation as directed by answer file. Set
   'q_fail_on_unsuccessful_master_lookup' to 'n' if installation
   should continue despite communication failures.

I haven't been able to use oscar sucessfully from windows. Am i missing something in my config?

[arubdesu]

Getting the same error with a non-similar setup(vmware fusion provide on Mac). When ssh'ing in to the agent it can ping, but can't telnet to the port of the master, because it can't route. My issues therefore seems inter-vm routing related... I know that doesn't help, but it may point in a direction.

[dec-pupop]

@arubdesu I had the exact same issue, I can't figure out why this is happening either.
To me the hostfiles seem to correct. although I did find when I logged onto master and ran puppet agent -t it failed until I removed one of the entries I think it was localhost, I will revert back with full information later. so maybe there is some issue there.

[cdenneen]

is your master still running.
I'm noticing 3.7.2 my 8140 dies after few minutes...
do netstat on the master to see if 8140 is still listening.

/etc/init.d/pe-puppetserver
pe-puppetserver dead but pid file exists

[dec-pupop]

@cdenneen
here are my outputs :

[vagrant@master ~]$ netstat | grep 8140
[vagrant@master ~]$

/etc/init.d/pe-puppet status
puppet (pid 11731) is running...

@arubdesu
[vagrant@master ~]$ puppet agent -t
Info: Creating a new SSL key for master.corp.emc.com
Error: Could not request certificate: getaddrinfo: Name or service not known
Exiting; failed to retrieve certificate and waitforcert is disabled

after removing '127.0.1.1 master' from hostfile on master it works.

[cdenneen]

@dec-pupop
pe-puppet is the agent not the server
pe-puppetserver status shows dead?
Your server is most likely dying too but nothing in pe-puppetserver.log

[dec-pupop]

ok ya I meant to take status of server. on master I see :

pe-httpd.worker (pid 18424) is running...

pe-puppetserver I don't think is in the version I am using, that comes in newer version right?

[dec-pupop]

any updates on this?

[adrienthebo]

@dec-pupop the original issue report indicated that you were having DNS resolution issues during installation, but the second latest comment indicates you have a PE agent installed but it can't ping the master. So to be clear, the issue you're working on is getting the agent to be able to resolve the master hostname, correct?

[cdenneen]

@adrienthebo I've had similar issue as described above where the puppetserver dies with no error logged.

[dec-pupop]

@adrienthebo

I am still using the same configuration, each time I run vagrant up I get the following error :

?? Install puppet master? [y/N] n
?? Puppet master hostname to connect to? [Default: puppet] master

Stderr from the command:

!! ERROR: Puppet Master at 'master:8140' could not be reached.
Aborting installation as directed by answer file. Set
'q_fail_on_unsuccessful_master_lookup' to 'n' if installation
should continue despite communication failures.

master doesn't have any entry in /etc/hosts for first (agent). so can't ping to first but can by ip.
on first (agent) it can ping master by hostname

on master I can see :
tcp 0 0 :::8140 :::* LISTEN

and
pe-httpd.worker (pid 18702) is running...

but when i execute :
[vagrant@first ~]$ telnet master 8140
Trying 10.20.1.2...
telnet: connect to address 10.20.1.2: No route to host

so not sure what the issue is here.

[arubdesu]

I just moved over to puppet-debugging-kit and all is well. Only wished I'd heard about it sooner, it layers on top of Oscar and there was even an intro vid last puppetconf

[weilera]

Maybe the firewall on the master node is blocking the connections.
see : #31

[marsmensch]

Ran into the same issue with the following parameters:

Boxes used for puppet master and agent: puppetlabs/centos-6.6-64-nocm
Oracle VM Virtualbox: 5.0.0
Vagrant: 1.7.3
Oscar: 0.4.2
vagrant-config_builder: 0.14.0
vagrant-hosts: 2.5.0
vagrant-pe_build: 0.10.3    

Manual fix:

$ vagrant ssh $YOUR_MASTER_HOST
sudo service iptables stop
exit
Then provision first node
$ vagrant provision $YOUR_MASTER_HOST

The following change to the master definition in vms.yaml worked also for me (moved firewall change to top):
https://gist.github.com/marsmensch/4903f6631da0eb391a04

[TJM]

FYI, this (like several other issues) is most likely that the puppet master has a firewall blocking port 8140 (as it says to check for in the output), but that scrolls by. It should probably manage a firewall rule or disable firewall.