The cluster is waiting for 10800s

[mhhli]

* I isntall sahara via kolla-ansible. And I open the dvr function.
  When I try to create a cluster via sahara. The status will be waiting for 10800s and get error last.
  Last I find in sahara/utils/ssh_remote.py (Line723), there is a command
      "proxy_command = (
                'ip netns exec qrouter-{router_id} nc {host} {port}')"

  But if I open the dvr, there will be a snat router on network node(my is also controller node).
  I try to change the "qrouter" to "snat" ,then I can create the cluster.
  So, I think whether can fixed the bug and I don't need to modify the ssh_remote.py when I reinstall sahara or I close
  the dvr function.

Thanks

[Luigi Toscano (ltoscano)]

I don't know how DVR works or what does change in the network setting, but you can redefine proxy_command from sahara.conf. Check sahara/utils/remote.py:

    cfg.StrOpt('proxy_command', default='',
               help='Proxy command used to connect to instances. If set, this '
               'command should open a netcat socket, that Sahara will use for '
               'SSH and HTTP connections. Use {host} and {port} to describe '
               'the destination. Other available keywords: {tenant_id}, '
               '{network_id}, {router_id}.'),

[mhhli (mssn)]

Thanks.

[mhhli (mssn)]

Now, I find if I use haproxy or I deployment network node on other computer which is no sahara_engine or sahara_api, this error will be hadppen, I think the command 'ip netns exec qrouter-fakerouter nc 10.0.0.2 22' will not exec, because there is no ip netns on the computer which sahara service run.
I find not only ssh but also http will exec the command(http port is 8080). but if I use the qdhcp namespace, I can exec the command correctly.
Could someone fix it ?

[Luigi Toscano (ltoscano)]

Please comment on storyboard, we don't use launchpad anymore:
https://storyboard.openstack.org/#!/story/1741456