Discovering OS Bugs and Using Snapshots

I’m running Debian/Unstable on an EeePC 701, I’ve got an SD card for /home etc but the root filesystem is on the internal 4G flash storage which doesn’t have much spare space (I’ve got a full software development environment, GCC, debuggers, etc as well as running KDE4). On some of my systems I’ve started the practice of having two root filesystem installs, modern disks are big enough that usually it’s difficult to use all the space, and even if you do use most of the space the use of a second root filesystem only takes a fraction of a percent of the available space.

Today I discovered a problem with my EeePC, I had upgraded to the latest Unstable packages a few days ago and now when I run X programs the screen flickers really badly every time it’s updated. Pressing a key in a terminal window makes the screen shake, watching a video with mplayer makes it shake constantly to such a degree that it’s not usable. If that problem occurred on a system with a second root filesystem I could upgrade the other a few packages at a time to try and discover the root cause. But without the space for a second root filesystem this isn’t an option.

I hope that Btrfs [1] becomes ready for serious use soon, it seems that the btrfs snapshot facility might make it possible for me to preserve the old version in a bootable form before upgrading my EeePC (although even then disk space would be tight).

So I guess I now need to test different versions of the X related packages in a chroot environment to track this bug down. Sigh.

7 comments to Discovering OS Bugs and Using Snapshots

  • Michael Goetze

    I don’t quite understand what features you expect from Btrfs that LVM doesn’t already offer you today?

  • Fedora is already working on this, they claim they’re 80% of the way there already:


  • vvill

    I’ve been running xserver in sid with 4 xserver packages for maybe 6 months.
    And I only use xorg.conf to try new available options.
    This set of packages is easily downgraded to the next highest working verions.
    Obviously, if you have the xorg metapackage installed it ‘s not so easy to do this trick. Plus I start from a base install in install x as lite as I can usually as well.
    I do have a sid KDE install I will take a look at eventually, but I was hoping to waiting untill X in sid gets a bit more stable again..
    I have the issue of “no visable text on the console after exiting an xsession”
    This happens for me with both xserver-xorg versions 1:7.5+1 and 1:7.5+2
    xserver-xorg-core 2:1.7.4-2
    xserver-xorg-input-evdev 1:2.3.2-3
    xserver-xorg-video-intel 2:2.9.1-2

    and a set that works
    xserver-xorg 1:7.4+3
    xserver-xorg-core 2:1.6.2-4
    xserver-xorg-input-evdev 1:2.2.2-1
    xserver-xorg-video-intel 2:2.8.0-1

    I’m starting to wonder if the last two upgrades of grubpc caused the console framebuffer issue I’m seeing after exiting xsessions.

    Sorry for any chatter, and checking out the x strike force mailing list
    I won’t add to the maylay unless I have something to lessen their workload.

  • Daniel Baumann

    Are you aware that you can install live-initramfs and aufs-modules (or kernel >= 2.6.32) on a non-live system, and add ‘boot=live plainroot’ as boot parameter (you might want to add it as a third boot entry in your bootloader configuration, besides the regular and the single user mode), and your regular system becomes a live system without persistency.

    provided you have enough ram, you can easily do a dist-upgrade and see if it works, if it doesn’t, reboot and all changes are gone.

    not as elaborate as btrfs will be at some point, though.

  • Anonymous

    By any chance do you have Intel graphics?

    Did you upgrade to the latest kernel and X driver?

    The latest bits include the work to downclock the LVDS when the image on the screen doesn’t change, to save power. As it turns out, a small handful of systems misbehave if Linux does this, with the symptoms you described. For that reason, 2.6.33-rc5 turns LVDS downclocking back off by default.

  • etbe

    Michael: I guess that LVM could theoretically do this, but I don’t expect it to be nearly as space efficient. Also if you have a small block device and you want to use LVM in that way then you lose the ability to use all the space. I could make a 3.6G root LV in the 4G flash storage to allow 400M for snapshots, but that would always reduce my space available to 3.6G max.

    vvill: I’m also wondering if it’s related to the framebuffer, but as I’ve purged the older kernels to save space it’s a little difficult to check.

    Daniel: That’s an interesting feature. Although with only 512M of RAM I will be a little limited in that regard on my EeePC. It will probably work really well for bigger systems.

    Anon: Intel Corporation Mobile 915GM/GMS/910GL Express Graphics Controller (rev 04)

    Thanks for the information, I will eagerly await the next version of the kernel. For the moment the primary task of my EeePC (fixing problems on servers in response to SMS from Nagios) can be performed in text mode.

  • vvill

    I like you have more than one install of sid on my machine… I get the same behavior from xserver using both kernels 2.6.30-2-686 and 2.6.32-trunk-686
    The reason I’m leaning towards grub-pc for “my” issue with xserver, Is that I lost “visible console text with the xserver-xorg 1:7.5 +1 upgrade
    than got the console back for almost 1 day. than after upgrading grub-pc later that day,lost the console again.. I admit, I have no clue about what the 18 ? new lines in /etc/grub.d/00_header are for. And at one point I tried to back out every package I’d upgraded to see if I could locate the offending package or file.
    Than console text was visible again and I upgraded all three installs, even the one I was using to test, which was starting to get a bit messy..

    This thread also seems to point to the framebuffer and includes LVDS as described in comment #5
    title [PATCH] fb/intelfb: Do not depend on EMBEDDED