Advice - bad RAM

[Chris Mohler]

On Sat, Sep 19, 2009 at 8:45 PM, Andrew Farris <flyindragon1 at aol.com> wrote:
> On Sat, 2009-09-19 at 20:20 -0500, Chris Mohler wrote:
> [snip]
>> I know that there is (or was) a kernel-level program for
>> "sitting on" bad addresses - anyone used this?  Also - anyone ever
>> have false positives with memtest86?
>>
>> I'm also probably hunting for new RAM - this machine is a Dell Vostro
>> 1500 laptop.
>
> if you need new RAM for your lappy, check ouy www.newegg.com. they've
> even got a RAM searcher where you can find RAM for your specific model
> of computer, or you can search by any other criteria under the sun just
> about.
>
> I have had false positives memtesting RAM when it gets really hot
> (usually happens after ~5 full passes for desktops w/o proper RAM
> cooling)
>
> I'd suggest ventilating your laptop very well (i.e. sit it on a cooling
> pad, or something) let the thing cool off for a while (turned off), then
> turn it on w/ the cooling pad, and test the RAM. If you get the errors
> again, then it's likely that it's bad.
>
> as for a kernel program for sitting on bad RAM addresses, i've never
> heard of it, but I'd be nervous about that anyway, because if your RAM
> starts failing now, it could start failing worse on you later, and it
> would be better to stomp out that possibility now while you're
> prepared.

Thanks for your thoughts.  I'm using the memmap kernel param to sit on
two large, error-prone chunks of RAM for now.  I'm going to try and
get some in-town tomorrow if at all possible.

But before that - I will go ahead and shut it down for  long while and
see if memtest still finds errors on a (really) cold reboot.

Chris