Re: [PATCH RFC 0/2] add nproc cgroup subsystem

[Tejun Heo <tj@kernel.org>]

Hello, Austin.

On Fri, Feb 27, 2015 at 01:49:53PM -0500, Austin S Hemmelgarn wrote:
> As far as being trivial to achieve, I'm assuming you are referring to rlimit
> and PAM's limits module, both of which have their own issues. Using
> pam_limits.so to limit processes isn't trivial because it requires calling
> through PAM to begin with, which almost no software that isn't login related
> does, and rlimits are tricky to set up properly with the granularity that
> having a cgroup would provide.
...
> PID's are a fundamental resource, you run out and it's an only marginally
> better situation than OOM, namely, if you don't already have a shell open
> which has kill builtin (because you can't fork), or have some other reliable
> way to terminate processes without forking, you are stuck either waiting for
> the problem to resolve itself, or have to reset the system.

Right, this is an a lot more valid argument.  Currently, we're capping
max pid at 4M which translates to some tens of gigs of memory which
isn't a crazy amount on modern machines.  The hard(er) barrier would
be around 2^30 (2^29 from futex side, apparently) which would also be
reacheable on configurations w/ terabytes of memory.

I'll think more about it and get back.

Thanks a lot.

-- 
tejun