AW: kill (work-around) does not wait. Was: Sending Signal to remote process

[Fiedler Roman]

> Von: openssh-unix-dev [mailto:openssh-unix-dev-
> 
> Am 24.01.2018 um 04:56 schrieb Carlos Morel Gómez La Roque:
> > Well, it would be nice to do
> >
> > ssh remote print-signal.py
> >
> > and then
> >
> > ssh remote pkill print-signal.py
> >
> 
> Yes, I could work around this by killing the remote command myself.
> 
> But "kill" or "pkill" have one big disadvantage in this case:
> ...

What about that (untested)?

* Start the remote process in a screen session with custom control key bindings?
* When termination is needed, send the "screen detach sequence" to remote
* You drop to shell outside screen, send signal via screen descriptor (no pkill hell) to inside process
* If process should not terminate by that signal (e.g. HUP to reload config), reattach to the screen again

Effect: killing works nice, for signal + resume, screen flickering would be seen.

LG R
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4814 bytes
Desc: not available
URL: <http://lists.mindrot.org/pipermail/openssh-unix-dev/attachments/20180124/3865f2de/attachment.p7s>