Mon Apr 19 01:35:04 CEST 2010
>From the way you have described it, it seems like all the processes
have been explicitly linked to the root process. This does not
automatically mean that they are linked to each other, they are only
linked to the root process. If one of them crashes then the exit
signal goes to the root process. If the root process is trapping exits
and does not itself crash then the exist signal will not be
transmitted to the other processes which will then continue.
Is this what you meant?
On 14 April 2010 19:17, John Erickson <> wrote:
> Hello, in this case I don't think there is any normal termination at all,
> just a network delay that causes the other nodes to think this node is down.
> Is there any way for the other nodes to die in this case?
> On Mon, Apr 12, 2010 at 10:50 PM, Rapsey <> wrote:
>> If a process dies with exit signal normal (like if the process function
>> stops recursion) the linked processes will not die. If there is any other
>> exit signal, linked processes will die as well. You can call exit(stop) or
>> something when you want to kill all processes as well.
>> On Tue, Apr 13, 2010 at 12:10 AM, John Erickson <> wrote:
>> > Hello, I have created a bunch of processes and called link() on them from
>> > root process. I thought this would mean that when one of them died, they
>> > would all die. However, I see in my log file one of the processes is
>> > responding' but the others run along oblivious. Is there a way to have
>> > them
>> > all die when this happens?
More information about the erlang-questions