This changes the Channel#wait call to a loop that checks
that the timeout has not been exceeded. If the timeout
has been exceeded, an SshTimeout error (subclass of SshError)
is raised.
This should catch any shell commands that take too long to
execute, as well as any hangs in the SSH connections, which
should prevent most current causes of apoptosis.
I'm including the timeout in the Clog emit, since it may
be useful.
This switches the newly added `sudo timeout 10s` in VM nexus
command with a `timeout: 10` argument. We can consider
expanding the use of explicit timeouts to other commands,
because the default timeout is fairly long (2-10 seconds less
than the apoptosis timeout).
This removes the SSH Channel#wait call inside the Channel#exec
block. Session#open_channel returns the same channel it yields,
and Channel#exec yields the receiver, so previously, we were waiting
twice on the same channel. This removes the duplicated wait inside
the block, as the wait outside the block will always be called.
7b0bad25ed