Re: [uml-user] swapoff killed by OOM despite there is enough RAM available

[richard -r. w. <ric...@gm...>]

On Wed, Nov 30, 2011 at 3:12 PM, Dmitry Panov <dmi...@ya...> wrote:
> On 30/11/2011 14:00, richard -rw- weinberger wrote:
>>
>> On Tue, Nov 29, 2011 at 6:39 PM, Dmitry Panov<dmi...@ya...>
>>  wrote:
>>>
>>> Hi guys,
>>>
>>> I'm having an issue with swapoff which gets killed by OOM. I was able to
>>> reproduce it with UML guests running different kernel versions, starting
>>> from 2.6.26 to  3.1.3. The only difference is with latest kernels
>>> sometimes
>>> it crashes instead of being killed. I haven't seen it on 'native' kernels
>>> so
>>> I guess it's UML specific.
>
>
>>>
>>> Can anybody reproduce this?
>>>
>> Yes (sort of).
>> OOM does not happen but swapoff() seems to run forever.
>> It's looks like the kernel is unable to unuse swapped pages.
>>
> With later kernels I had the same effect, swapoff running forever. Earlier
> versions get it killed by OOM which is actually better as it at least allows
> the machine to shutdown.
>
> Any ideas what could be the problem?
>

I have to talk to some mm hackers.
I'm not sure how this issue is related to UML.

Maybe UML triggers a mm-bug.
Or (most likely) mm triggers a bug in UML.

-- 
Thanks,
//richard