Defending against fork bombs
Defending against fork bombs
Posted May 5, 2005 9:48 UTC (Thu) by NAR (subscriber, #1313)Parent article: Defending against fork bombs
Then, rather than killing the second chosen victim directly, the OOM killer goes after the ancestor - and all of its children.
According to the pstree, the common ancestor of all my applications started from the window manager's toolbar (including applications started from xterms started from the toolbar) is the window manager itself. So if two of these applications manage to trigger the OOM killer, there is a chance that all of my running applications will be killed. I don't know the probability of this scenario (probably pretty low) but it still doesn't sound to be a good idea for a desktop. And what happens when the fork bomb daemonizes itself (i.e. it's direct ancestor will be the init process)? I don't think it's that hard to create a fork bomb like that...
Posted May 5, 2005 15:20 UTC (Thu)
by proski (subscriber, #104)
[Link]
I guess that's why "biological" parents need to be tracked.
And what happens when the fork bomb daemonizes itself (i.e. it's direct ancestor will be the init process)?Defending against fork bombs