Re:Killing processes that don't want to die
by CamperBob2 on Monday September 24, 2018 @10:21PM
(#18063080)
How is that not a fork bomb? ("The author disagrees with this characterization, which was added by an editor late in the publication process.")
[ Parent
| on HN ]
Re:Killing processes that don't want to die
by CGamesPlay on Monday September 24, 2018 @10:32PM
(#18063120)
Because the number of live processes never exceeds 2, and fork bombs typically refer to DoS attacks by exhausting the PID space.
Re:Killing processes that don't want to die
by nneonneo on Monday September 24, 2018 @10:28PM
(#18063104)
The parent always exits (fork() returns the nonzero child pid in the parent, and 0 in the child), meaning that at any point there’s only one process that is actively forking. A typical fork bomb, on the other hand, floods the system with processes that are all attempting to call fork, resulting in an exponentially growing set of processes until the system runs out of resources.
Re:Killing processes that don't want to die
by userbinator on Monday September 24, 2018 @11:34PM
(#18063379)
I propose the term "forkworm", or perhaps "forxolotl"[1] for these processes that continuously reproduce but don't grow exponentially.
Re:Killing processes that don't want to die
by caf on Tuesday September 25, 2018 @12:19AM
(#18063566)
It already has a name - processes like this were historically called "comets".
Re:Killing processes that don't want to die
by OskarS on Tuesday September 25, 2018 @03:26AM
(#18064267)
This is a fork bomb:
while (1) {
fork();
}
This will fork forever, creating children that fork forever, exhausting the system's resources. The example in the article only ever has at most 2 processes, and they don't exhaust anything.