SCIENTIFIC-LINUX-USERS Archives

December 2011

SCIENTIFIC-LINUX-USERS@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Jean-Paul Chaput <[log in to unmask]>
Reply To:
Jean-Paul Chaput <[log in to unmask]>
Date:
Wed, 21 Dec 2011 00:48:39 +0100
Content-Type:
text/plain
Parts/Attachments:
text/plain (56 lines)
Hello Scientific Linux Users,


I'm using SL 6 on my laptop and since the migration to SL6 it has
become more and more unstable. The fact that failures arose
more or less randomly in throughout all the system finally hinted
me toward a memory corruption linked with hibernation (suspend to
disk).

After googling a little I found out that is was a known and corrected
bug:
  http://us.generation-nt.com/answer/patch-pm-hibernate-fix-memory-corruption-related-swap-help-201280532.html

But unfortunatly, it starts with 2.6.38 and TUV seems not to have
backported it yet to it's 2.6.32.

Using kernel-ml 2.6.39 from elrepo solve the problem.

Then, I ran into another one: after less than 10 hibernation cycles,
the laptops reboot instead of resume. That is a known bug of the
xorg-x11-drv-intel (I use a core i5), and you have to upgrade
to the 2.16.0 version (also available through elrepo) and put
a "nomodeset" into your kernel parameter list (grub.conf).
This is an issue related to the new KMS feature.

To completly suppress the "reboot" effect you also need to upgrade
pm-utils to a version at least 1.2.6 which handle correctly KMS
detection at resume (I use 1.4.1 from Fedora 16).

I'm wondering if anyone has also noticed the problem?
(or maybe I'm completely mistaken)

Note 1: The memory corruption is a relatively slow process, it shows
        only if you uses swap. With 2Gb of RAM it took between 3 to
        7 days to trip on something vital for the system (for example
        the glibc). And more than a week with 4Gb of RAM.

Note 2: From what I read on various forums, the mixing of KMS problems and
        memory corruption seems have mislead people toward a pure
        X11 problem.

Note 3: It doesn't affect the suspend to disk.


In the hope it will be useful.

-- 
      .-.     J e a n - P a u l   C h a p u t  /  Administrateur Systeme
      /v\     [log in to unmask]
    /(___)\   work: (33) 01.44.27.53.99              
     ^^ ^^    cell:      06.66.25.35.55   home: 01.47.46.01.31

    U P M C   Universite Pierre & Marie Curie
    L I P 6   Laboratoire d'Informatique de Paris VI
    S o C     System On Chip

ATOM RSS1 RSS2