Re: [discussion] Swap overcommitment recovery

david (david@kalifornia.com)
Mon, 17 Aug 1998 11:03:55 -0700


Reply to mail from Andi Kleen about [discussion] Swap overcommitment recovery
> The problem usually then is that the X server is blocking in swapping,
> and thus can't handle the mouse pointer. The real fix for that is a
> multi threaded X server, not some crazy kernel hacks.

Actually, gpm at the console as well as typing at the console is affected.
Everything is affected. I have had this happen many more times on a server
without X running because of one reason or another. The recent Apache DoS
attack comes to mind. Resource limits aren't all that useful when the
server is designed specifically as a web server and *should* get 99% of
the memory. Limits are in place to try and keep this to a minimum, but
the odd occurance does happen.

Back to the story. Userland control becomes fairly non-existant. Whether
this is my personal workstation or a server, the situation needs to be
resolved immediately and userland is way too slow to do this.

I don't consider waiting 10 minutes for ctrl-alt-del to take effect a
proper solution. The kernel overcommits memory which is a necessary and
desirable feature; the kernel needs to help me recover. The last entry on
the list, putting everyone to sleep and waking up the userland daemon to
deal with things is IMO a fair solution.

-d

-- 
Look, look, see Windows 98.  Buy, lemmings, buy!   
(c) 1998 David Ford.  Redistribution via the Microsoft Network is prohibited.
 for linux-kernel: please read linux/Documentation/* before posting problems

- 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.altern.org/andrebalsa/doc/lkml-faq.html