Bug#212117: ssh: init stop/restart action should use --retry

[Colin Watson <cjwatson@debian.org>]

tags 212117 pending
thanks

On Mon, Sep 22, 2003 at 03:24:53PM +1000, herbert@gondor.apana.org.au wrote:
> On one of my slower machines I lost sshd completely after upgrading
> to the latest security update (the unstable version has the same bug
> as well).  The reason is that the init script tried to start before
> the old sshd is completely stopped:
> 
> Sep 22 05:03:56 host sshd[18705]: Received signal 15; terminating.
> Sep 22 05:03:58 host sshd[1759]: error: Bind to port 22 on 0.0.0.0 failed: Address already in use.
> Sep 22 05:04:16 host sshd[1759]: fatal: Cannot bind any address.
> 
> This is probably because the 2-second delay in the init script is
> not sufficient.
> 
> Rather than extending the delay, please call start-stop-daemon --stop
> with --retry so that it does not return until the daemon is truly dead.

Thanks for the suggestion; I've fixed this in CVS. As a bonus,
restarting is now quick if the daemon stops quickly.

Cheers,

-- 
Colin Watson                                  [cjwatson@flatline.org.uk]