[ale] After 15 years, Nohup is sttll broken???

[neal at mnopltd.com]

Well, I apply for Medicare in 6 months, so...

Yes, "screen" or "tmux" is a workaround.  I'm doing this server upgrade 
for admin staff that will have to handle it day-to-day.  IMHO, "screen" 
is a bit confusing for them to follow.   One cannot write down rote 
commands for figuring out which screen session you left if more than one 
is in play.

It all still begs the basic question: the man page for Nohup states the 
following as its description:   "Run COMMAND, ignoring hangup signals."  
  Apparently it fails to do its one and only purpose in life.

Yes, we have gone back to use "at" instead, which DOES work no matter 
how the tty/ssh session dies.   But I still find it curious.

Regards,

Neal


On 2019-07-31 09:28, Jerald Sheets wrote:
>> On Jul 31, 2019, at 10:19 AM, Chris Fowler via Ale <ale at ale.org>
>> wrote:
> 
>> tmux.  Start the tmux session and run the job in a tmux pane.  When
>> you get home SSH back to the server and attach to the tmux session.
>> I use this in development so I can occasionally check up on build
>> status via JuiceSSH on Android.
> 
> THIS.  I was going to be old and mutter something about screen, but
> tmux is the new hotness.
> 
> Just because you’re not setting keepalive does not mean SSH will not
> intrinsically keep the connection alive or for some uninstrumented
> reason clamp the session.  Or a network hiccup.  Or OOMkiller being a
> whiny bitch.
> 
> Literally any number of things can be hanging up that session for you
> irrespective of your nohup.
> 
> Set the keepalive, or use tmux.
> 
> —j