Henchman should timeout if the remote machine didn't respond for long time.

[baskaran-md]

One of the instance in inventory was bad such that its up but responding very slowly for ssh and other remote requests.. In this case henchman was waiting indefinitely for the command to complete. It has to detect those bad instances and timeout accordingly.

[jlin21]

How about setting standard time out period for each task. And modules can override that time delay

[sudharsh]

Modules shouldn't do it. Let it be at a task level if we need such a timeout.

[jlin21]

There would be a timeout field for each task then? Should tasks by default have a standard timeout value?

[sudharsh]

yeah, as defined by the transport. It's transport's responsibility

[jlin21]

If a task times out. Should it error out or just mark that state as 'skipped' or 'unreachable'

[sudharsh]

Timeout should be treated as unreachable. More I think of it, more I feel we should standardise on the states On 23-Nov-2015 4:04 pm, "jlin21" [email protected] wrote:

If a task times out. Should it error out or just mark that state as 'skipped' or 'unreachable'

— Reply to this email directly or view it on GitHub https://github.com/apigee/henchman/issues/82#issuecomment-159109370.

[jlin21]

Yeah we have a few redudant states such as ok, changed, and failure, error.

[jlin21]

Actually we should probably have a state named timeout since unreachable means henchman cannot even reach the remote system

[jlin21]

https://github.com/golang/go/wiki/Timeouts