Hi!
On Thu, Apr 02, 2009 at 12:54:45PM -0600, RB wrote:
> PaX. The memory management changes it provides in combination with
> the memory management changes introduced in 2.6.28 could well
> exacerbate existing issues in a RAM module that weren't being
> triggered previously.
Yeah, that's true. But it's a production server, and it's hard to switch
it off for ~6 hours while memtest86 will check RAM.
Also, nature of this bug make it ease to debug it, because it's happens
every time. I'd prefer to find a way to trace PaX code to find out what
exactly goes wrong, why it kill process N1. If PaX team provide a patch
which add some debugging output for PaX, or something else - to make this
investigation possible - I'll be happy to try it.
--
WBR, Alex.
