Timeout using ssh config ProxyCommand

There is, what I would consider, a bug in the way ssh config files are used to proxy.

Normally your ssh config allows you to do this:

Host *
    ServerAliveInterval 5

Host *.domain
    ProxyCommand ssh -W %h:%p domain_proxy

Host my_host.domain
    HostName 10.10.10.10

when you ssh to my_host.domain, it will apply the ServerAliveInterval from Host *, the ProxyCommand from *.domain, and the IP address from my_Host.domain. like so:

> ssh -vvvv my_host.domain
OpenSSH_7.2p2 Ubuntu-4ubuntu1, OpenSSL 1.0.2g-fips  1 Mar 2016
debug1: Reading configuration data ssh_config
debug1: ssh_config line 1: Applying options for *
debug1: ssh_config line 17: Applying options for *.domain
debug1: ssh_config line 45: Applying options for my_host.domain*
debug1: Executing proxy command: exec ssh -W 10.10.10.10:22 domain_proxy
...

Netmiko, on the other hand, will not find the ProxyCommand under *.domain, you must set all of the options under the Host its self like so:

Host my_host.domain
    HostName 10.10.10.10
    ProxyCommand ssh -W %h:%p domain_proxy
    ServerAliveInterval 5

If you must use a proxy to connect, the first ssh config, without ProxyCommand specified under the host, will give this error:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/lbernard/miniconda3/envs/netmiko/lib/python2.7/site-packages/netmiko/ssh_dispatcher.py", line 84, in ConnectHandler
    return ConnectionClass(*args, **kwargs)
  File "/home/lbernard/miniconda3/envs/netmiko/lib/python2.7/site-packages/netmiko/base_connection.py", line 68, in __init__
    self.establish_connection(verbose=verbose, use_keys=use_keys, key_file=key_file)
  File "/home/lbernard/miniconda3/envs/netmiko/lib/python2.7/site-packages/netmiko/base_connection.py", line 169, in establish_connection
    raise NetMikoTimeoutException(msg)
netmiko.ssh_exception.NetMikoTimeoutException: Connection to device timed-out: cisco_ios my_host.domain:22

[lundybernard]

Thanks @ktbyers that example will work perfectly.

The only thing that’s not making a lot of sense to me, is the setting the host record to <ip>:<port>, since the record already contains the IP under ‘HostName’, and the ‘Port’.

as in your example:

host 10.10.10.27:9822           # Juniper SRX
  user pyclass
  hostname 10.10.10.27
  Port 9822
  ProxyCommand ssh jumphost nc %h %p

It feels redundant, and I have always seen ssh config files set the host record to the FQDN, or a convenient alias. I would expect this config to look like:

host Juniper.SRX
  user pyclass
  hostname 10.10.10.27
  Port 9822
  ProxyCommand ssh jumphost nc %h %p

or some similar naming scheme.

One thing I have not explored is how ssh config interacts with records in /etc/hosts, so i’m unsure if setting host=<ip>:<port> is due to that consideration. Is there some other commonly used application that is driving the ip:port naming scheme?

Regardless, I will try to get a pull request together to satisfy both scenarios. I am not terribly familiar with pytest (I use unittest on my projects). Could you direct me to an existing test file that I should add tests to, or an appropriate new file name for the tests, that will be inline with your schema?

[ktbyers]

https://github.com/ktbyers/netmiko/commit/24d39ce28484af0b4465f53c55e1ed0342f92c30