Re: Overcommitable memory??

From: Jesse Pollard (pollard@tomcat.admin.navo.hpc.mil)
Date: Tue Mar 21 2000 - 15:41:05 EST


Horst von Brand <vonbrand@pincoya.inf.utfsm.cl>:
> David Whysong <dwhysong@physics.ucsb.edu> said:
>
> [...]
>
> > Preventing system OOM using resource limits is equivalent to disabling
> > overcommit. You have to restrict each of N users to 1/N of the total
> > system memory.
>
> So you have 500 accounts, each one gets to use 1/500th of the (be very
> generous!) 2Gb VM. What if a batch of 50 new accounts shows up? Everybody
> gets the quota cut down 10%?

I believe the "N" referrs to N concurrent users.

> There are usually 10 active accounts, each one gets 200Mb. What if the 11th
> logs in? Start killing off processes of the people that were using exactly
> their share of the memory ATM?

The 11th gets an "insufficient resourcess currently available for login"
message.

> Resource limits per user are orthogonal to overcommit. You could very well
> limit Jane Random User to 300Mb, as long as the system is able to provide
> it. You can overcommit or not, but in no case can Jane use more than her
> limit.

-------------------------------------------------------------------------
Jesse I Pollard, II
Email: pollard@navo.hpc.mil

Any opinions expressed are solely my own.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Thu Mar 23 2000 - 21:00:34 EST