[olpc-software] graceful handling of out-of-memory conditions

[David Zeuthen]

On Tue, 2006-03-28 at 10:44 -0500, Alan Cox wrote:
> > I really think we don't want to kill them; we want a rock solid base of
> > services we can always assume are there; sure if HAL or NM suddenly
> > needs to allocated some memory one of the "applications" (e.g. Firefox,
> > Abiword) will have to pay for it by getting shot by the OOM killer...
> > Yes?
> 
> Do you want the box to crash or a server to restart. Your choice. I'd rather
> that in the 'I've killed all the apps and its not helped' case we killed the
> likely offending service.

Maybe. Or maybe not. I'm undecided.

Maybe I would just reboot the box? It's the same net result anyway! I
mean.. in the situation where we don't have room for "desktop shell"
processes even when no application is running we've already lost. By
rebooting at least we start from clean state. Btw, this only happens if
something is totally wrong as under normal circumstances we would have
lots of free RAM (say 80MB or so on a 128MB system) when only "desktop
shell" processes are running. Of course, if this is 80MB or whatever,
we'll have to see and wait...

> > I think we would want to never refuse allocations to the "desktop shell"
> > components (e.g. WM, panel, HAL etc.) as long as we have
> > "applications" (e.g. Firefox, Abiword) to shoot. Doesn't this make
> > sense?
> 
> Conceptually yes. I'll have a think about implementation issues.

Sounds good; thanks for looking into this!

    David