Set ssh timeout to be resilient for poor network connections

[mark-dce]

ISSUE
As configured in the defaults, the ssh session timeout may cause the session to be more likely to die unexpectedly on poor network connections. Since there are no retries, a very short network dropout at just the right time can disconnect you even when you're actively editing a file...

CURRENT SETTINGS
ClientAliveInterval 600
ClientAliveCountMax 0

PROPOSED SETTINGS
ClientAliveInterval 120
ClientAliveCountMax 5

In the proposed settings, the client would have to fail at five consecutive 2-minute checks (i.e. 10 minutes), but would remain alive if the system provides a successful response in any one of the tries. In the original, the session dies immediately on a single check at 10 minutes with no retries.

[acozine]

The defaults were set to meet the production spec for laevigata, with only the one variable available for override:

 # manage inactive ssh sessions
TCPKeepAlive no
ClientAliveInterval {{ ssh_timeout }}
ClientAliveCountMax 0

I agree that adding a variable for the retries is a good idea. Let me noodle on this for a bit.

[mark-dce]

We may want to think about backwards compatibility and deprecation notification - i.e.

Release X implements new ssh timeout settings and deprecates use without inactivity checks

• Please set ssh_timeout to your desired value in your config values, default is 0 (to match previous builds)

Release X+1 implements a 10 minute ssh inactivity timeout

• Please explicitly set your ssh_timeout to 0 if you wish to maintain previous behavior, default is 600 seconds.

[mark-dce]

Moved to https://github.com/curationexperts/ansible-samvera