Ok, what in the world is wrong with my system?

Discussion in '*nix Software' started by Red Squirrel, Jan 26, 2013.

  1. Red Squirrel

    Red Squirrel Lifer

    Joined:
    May 24, 2003
    Messages:
    36,925
    Likes Received:
    632
    I've been having so much trouble since I built this machine, Windows 7 had an issue where the monitors would randomly shut off for a couple seconds followed by a notice that the driver crashed. Got fed up and went with Linux in hopes things would be better, and I had been wanting to go with Linux anyway. Solved THAT problem but brought on many others. Running Xubuntu now (not sure how to check version but think it's still the latest)

    I'll start with the specs:

    - 12GB ram
    - Core i7 processor
    - 2x Nvidia GTX 560TI
    - Gigabyte X79 UD3 motherboard
    - SSD1: OCZ Vertex 3 (drive for Linux, primary used OS)
    - SSD2: Intel M4-CT128M4SSD2 (drive for Windows, rarely used)

    All the problems I've been experiencing so far:

    - If I try to use nouveau driver, screens constantly flicker like mad. The install process was quite brutal to go through since it did it then too. Any Linux live CD does not work properly on this system because of this insane flickering. Very hard on the eyes and just makes my blood boil.
    - Apps randomly crash with no warning. They'll just completely shut off and produce a "crash warning detected" notice. Sometimes it's daemons that I don't really notice that are gone, other times it's apps I have open.
    - Any music player randomly start to play just silence, or stop altogether, or crash
    - Cursor will get stuck in a way where I cannot click or do anything with it, but I can move it. Usually it will get stuck in the hand icon or other icon than the normal cursor. When this happens, I lose complete control of the keyboard so I cannot do anything. At one point I did not lose control of the keyboard so I was able to go to the console and close the offending app (usually Thunderbird or Firefox). But now when it happens I have no keyboard. Have to hit reset. The system itself is responsive when in this state as far as I know, it's just that I lose all ability to input so I cannot do anything with it.
    - Every time I reboot, my theme changes to some weird black theme. Not a standard one that I can actually pick.
    - my 3rd monitor runs in a separate X session (can't figure a way to get all 3 going at once without some very ugly workarounds) and often it will mess up. All sorts of weird artifacts on the screen etc. Need to reboot.

    And many other issues but those are probably the biggest I can think of now. I'm stating to think it's all related.

    Just now, I had minecraft start to get extremely laggy, then it shut off AND my music player (Gaydeleque) closed too! WTF is going on here?

    This is the output of dmesg.

    Out of memory?!?!?!I have 12 freaking gigabytes of ram, how can there not be enough memory for minecraft and a couple other apps? This particular issue is the first time it happens though. I'm at wits end here with this stupid system, wtf is wrong?

    I just want a system that stops acting stupid like this, what do I do? I really hate having to throw money at the problem but I'll do anything at this point. It's just so frustrating to deal with this crap.
     
    #1 Red Squirrel, Jan 26, 2013
    Last edited: Jan 26, 2013
  2. TheRyuu

    TheRyuu Diamond Member

    Joined:
    Dec 3, 2005
    Messages:
    5,406
    Likes Received:
    6
    This is random and probably not related to (most of) your problems but something I randomly noticed:
    Code:
    [ 7497.940127] [ 1476] 0 1476 3995 26 4 0 0 irqbalance
    Did you install it? Google says it should no longer be included as of 12.10 which is the latest which you (sort of) claim to have.

    irqbalance does not belong on a desktop machine.
    Code:
    sudo apt-get purge irqbalance
    (and reboot)

    And it should go without saying that the nouveau driver is no substitute for the official nvidia binary blob.
     
    #2 TheRyuu, Jan 26, 2013
    Last edited: Jan 26, 2013
  3. BrightCandle

    BrightCandle Diamond Member

    Joined:
    Mar 15, 2007
    Messages:
    4,763
    Likes Received:
    0
    You are running SLI, you need the nvidia binary drivers.
     
  4. Kiska

    Kiska Senior member

    Joined:
    Apr 4, 2012
    Messages:
    232
    Likes Received:
    3
    Or could be that OpenGL decided to screw up. And from what I can identify from the log you have SLI on so please follow.
    1. apt-get install nvidia-stable in recovery (sudo may be used to install stable nvidia binaries.
    2. reboot afterwards usually by invoking shutdown -r 0 (warning you may need to use sudo in the beginning) or you may hold down the power button until your system is shut down
    3. Check for errors or instabilities also if so mention them in this thread
    Italics identifies command to be used
    To get version please run the system-monitor. However I suspect version 12.10 is used. To do this find the program in system tools or by going to terminal and invoking command gnome-system-monitor. Also to start into recovery I believe is the second option on grub so please take your time in the system book up screen. Usually it has normal functionality and then the second option is your recovery mode, denoted by parenthesis. Also please have your Ethernet cable plugged in as the wireless WILL NOT work in recovery.
    James aka Kiska
    Your helpful 15 year old
     
  5. lxskllr

    lxskllr Lifer

    Joined:
    Nov 30, 2004
    Messages:
    44,665
    Likes Received:
    91
    If it were me, I'd strip the machine down, use minimal hardware, and see if it works like that. If it works, start adding parts until it doesn't.
     
  6. Red Squirrel

    Red Squirrel Lifer

    Joined:
    May 24, 2003
    Messages:
    36,925
    Likes Received:
    632
    No SLI. Just two cards. I originally had the connector but decided to remove it in case it's part of my problems. Should I still do the steps mentioned? The driver I'm using is NVIDIA Accelerated graphics driver (post-release updates) (version current-updates). Second option under additional drivers. If I pick the one on top, I lose ability to have more than 1 monitor as everything is grayed out in the configuration tool and it says it's not the right version. Should I try the betas, or is that asking for even more trouble?

    There's not really anything I can strip from the machine. The only thing I have is the 2 cards and I need them for the 3 monitors to work. The most I could do is go with just 2 monitors and remove one card but these problems happened even when I had just the one card as I had already tried going with just one, changing slot/card etc. I could try going with 4GB of ram and take the other sticks out though, guess I can start with that.

    Not sure where irqbalance came from, maybe it was a dependency for some other app so it got installed? What does it do?


    Either way, do these issues really sound video card / driver related? If yes I'll just bite the bullet and buy another 2 cards or see if I can find a card with 3 outputs, and that way I will rule out the hardware and the driver. Perhaps I should go ATI?
     
  7. ControlD

    ControlD Diamond Member

    Joined:
    Apr 25, 2005
    Messages:
    5,204
    Likes Received:
    14
    What size and type of power supply do you have running your rig? It certainly sounds like you have a real hardware issue if you are having problems in both Windows and Linux.
     
  8. ControlD

    ControlD Diamond Member

    Joined:
    Apr 25, 2005
    Messages:
    5,204
    Likes Received:
    14
    I am far from an expert at reading dmesg output, but it sure looks to me like you are getting a lot of "out of memory" errors for some reason. With 12GB of RAM that seems hard to believe, but it pops up several times.

    What other processes/services do you have running full time on your system? How is your swap partition set up? For 12GB of RAM I think you should have a swap partition of at least 24GB going by the old 2xRAM rule of thumb. I had some strange memory related errors recently when I upgraded my system from 4GB to 8GB of system RAM. Once I went and resized the /swap to 10GB all of those issues went away.
     
  9. lxskllr

    lxskllr Lifer

    Joined:
    Nov 30, 2004
    Messages:
    44,665
    Likes Received:
    91
    Regarding the swap, I think that would be a bandaid if it worked. It would be taking up slack from a memory leak or something. I run 1gb swap with 8gb ram, and I just setup a bootable thumb drive for my daughter, and it has no swap at all.

    I think it's hardware also, and the psu would be a good guess. That's why I suggested stripping the hardware. Not only would it be taking less power, it would help rule out other problems.

    Regarding not wanting to use a single monitor. Using a 7" display on a rock solid system would be better than 3 monitors on a crash fest. It could also get worse, and you could end up with file system corruption, or something that isn't easy to recover from. I'd try to fix it sooner rather than later.
     
  10. TheRyuu

    TheRyuu Diamond Member

    Joined:
    Dec 3, 2005
    Messages:
    5,406
    Likes Received:
    6
    It used to be installed by default on ubuntu I think, which was plain wrong since it doesn't belong on a desktop (so it was removed as of 12.10).

    From the google code page:
    It's for high throughput environments which don't care about latency (servers). So on a desktop machine where the ONLY thing we care about is latency we don't want (or need for that matter) irqbalance.

    Do you really want to solve this problem or just throw money at it until it hopefully goes away?

    Removing irqbalance is a start, the next step would be determining if there is a hardware problem or not. You might not like it for a while but you're going to have to use one monitor to figure out if there is a problem, I shouldn't have to say this but it will only be temporary (you seem to not want to do it).

    Try out the bare minimum needed to boot and use your system.
     
    #10 TheRyuu, Jan 26, 2013
    Last edited: Jan 26, 2013
  11. Red Squirrel

    Red Squirrel Lifer

    Joined:
    May 24, 2003
    Messages:
    36,925
    Likes Received:
    632
    PSU is 1000w OCZ Fatality.

    I only have 4GB for swap, do I really need to do double the ram? That's a large chunk out of the SSD drive but I suppose I could try that, or throw in a spindle drive.

    I just installed my monitoring software on this machine to monitor ram usage, so I will keep an eye to see if it goes into alarm. Wondering if I have something leaking memory? Though that should not cause all the other weird stuff going on or would it?

    I suppose it could be the PSU, but wouldn't it be more likely that it gives me issues while gaming? I sometimes play UT3 (In Windows) for a couple hours and never had any issues while gaming.

    Also am I better off leaving the SLI connector off, or should I put it back on? Even if I don't have SLI enabled right now should I still follow Kiska's steps?

    Irqbalance is also now removed. I just have to reboot.
     
    #11 Red Squirrel, Jan 26, 2013
    Last edited: Jan 26, 2013
  12. Red Squirrel

    Red Squirrel Lifer

    Joined:
    May 24, 2003
    Messages:
    36,925
    Likes Received:
    632
    Oh edited my original post, another issue I have is every time I reboot, my theme changes to some weird black theme. May also be worth mentioning that in Minecraft, occasionally my character will act like I'm holding down on a key. I need to esc out or tap a bunch of keys to make it stop. Wonder if it's something that's happening at the Linux environment level, as opposed to the game itself. I Googled it and it only happens to people in Linux. I call it the Toyota bug. usually causes my character to run off a cliff or into lava.
     
  13. TheRyuu

    TheRyuu Diamond Member

    Joined:
    Dec 3, 2005
    Messages:
    5,406
    Likes Received:
    6
    Well you still need to figure out if there's any hardware problems going on but you should also try the lowlatency kernel (don't forget the headers, nvidia driver needs them).

    The default generic kernel is just that, it's generic. But remember, we're on a desktop where the only thing we care about is latency (the differences are minimal, it's just tuned to provide lower latency by using real preemption (probably more important) and upping the kernel tick rate).
     
  14. Red Squirrel

    Red Squirrel Lifer

    Joined:
    May 24, 2003
    Messages:
    36,925
    Likes Received:
    632
    How would I go about doing that? I've never dared to mess with the kernel before. I do recall needing kernel headers to get vmware server to work on another system, and that was quite involved, but other than that never played much with the kernel.

    Also another issue I just remembered: my 3rd monitor runs in a separate X session (can't figure a way to get all 3 going at once without some very ugly workarounds) and often it will mess up. All sorts of weird artifacts on the screen etc. Need to reboot to fix it. The 3rd monitor is nearly unusable because of this. It will work for about an hour before it starts to mess up.

    Here's an example of what sometimes happens:

    [​IMG]

    Othertimes I have these lines that keep following the cursor around. If I bring the cursor on that screen I sometimes have to "ram" it into the side so I can get it back to the other 2 monitors.
     
  15. biodoc

    biodoc Diamond Member

    Joined:
    Dec 29, 2005
    Messages:
    4,602
    Likes Received:
    1
    Did you know you have Folding@Home SMP running on your computer? Looks like it's using 100% of 7 logical cores. If your system is not stable (overclocked?), Folding@Home will cause issues.
     
  16. TheRyuu

    TheRyuu Diamond Member

    Joined:
    Dec 3, 2005
    Messages:
    5,406
    Likes Received:
    6
    So have you figured out if it's a hardware problem (because it looks like one)?

    Does F@H not have a native linux binary?
     
  17. Jodell88

    Jodell88 Diamond Member

    Joined:
    Jan 29, 2007
    Messages:
    9,451
    Likes Received:
    27
    The Minecraft thing is with Minecraft I believe. If you run the game fullscreen you won't have that issue. If you can't turn 360 degrees press esc twice.
     
  18. Red Squirrel

    Red Squirrel Lifer

    Joined:
    May 24, 2003
    Messages:
    36,925
    Likes Received:
    632
    Yeah that's normal. I originally had it to 8 cores but decided to bring it down to 7 when the issues started to see if it does anything.

    Should I try disabling it completely? Did not figure it could cause any issues, but suppose it's worth a shot. Also I'm running stock, no overclocking. Too many issues to even start thinking about that.

    I don't know if it's hardware or not, that's what I'd like to figure out. Long shot here, but could it actually be my mouse? At least for that cursor stuck issue.

    As for Minecraft, tried going fullscreen but it tries to use both screens at once, then reverts back to the main, but the whole game is pushed sideways so I can only see one part. But if that's Minecraft specific then wont worry about it for now, I have enough more serious issues on my plate.
     
  19. TheRyuu

    TheRyuu Diamond Member

    Joined:
    Dec 3, 2005
    Messages:
    5,406
    Likes Received:
    6
    I'd start with the video card given that screen shot.

    Try one video card at a time, do all the normal stuff you would do (except on two monitors instead of 3) then swap and try again. Whichever one is driving the monitor you had in that screenshot, I'd try that one by itself first and see if you can still get any video related problems.
     
  20. Red Squirrel

    Red Squirrel Lifer

    Joined:
    May 24, 2003
    Messages:
    36,925
    Likes Received:
    632
    I tried this already before I got the 3rd monitor. I was running with just one card for a while, I even tried it in another slot, then also tried swapping the cards completely. (two of the same model). When I was in Windows I also tried running furmark on both cards, not a single issue. Is there an equivalent program in Linux? I'd like to try it again.

    Here's some more screenshots of weird artifacts.

    I had to take a picture of this one as it does not show up if I take a screenshot:

    [​IMG]

    Those two lines follow the cursor around. They sometimes go away. The missing text should say "open in new folder". If I pass the cursor over, it fixes it.

    This one is more weird than anything, this is on the main X session with the two monitors. The empty void (one monitor has higher res than the other) seems to have bits and pieces of the last boot.

    [​IMG]

    This is a void though, so not really sure what normally goes there, my guess is it should just be black. It's not a big deal on it's own, but just thought it might help diagnose the issue.
     
  21. ControlD

    ControlD Diamond Member

    Joined:
    Apr 25, 2005
    Messages:
    5,204
    Likes Received:
    14
    Have you tried booting your system with a live Linux DVD (any flavor) to see what happens? This should at least let you know if the problem is purely hardware as opposed to a software issue.
     
  22. Red Squirrel

    Red Squirrel Lifer

    Joined:
    May 24, 2003
    Messages:
    36,925
    Likes Received:
    632
    Because of the screen flickering live CD/DVDs are a very terrible experience to use on this computer. When I installed it, I pretty much had to speed through the install questions so I can let it do it's thing and go out of the room. It's that bad. Otherwise that's what I thought of doing as well.

    Come to think of it, I have not tried the latest Fedora Core. I think I will try this. That way it's a completely different distro architecture, and maybe things will be better. At this point I'm willing to switch distro just for my computer to work properly, and getting the triple display working as one X session would be nice as well... Maybe FC will be better, worth a shot. At least assuming it's not a hardware issue...
     
    #22 Red Squirrel, Jan 26, 2013
    Last edited: Jan 26, 2013
  23. Jodell88

    Jodell88 Diamond Member

    Joined:
    Jan 29, 2007
    Messages:
    9,451
    Likes Received:
    27
    Can you post your /var/log/Xorg.0.log ?
     
  24. Kiska

    Kiska Senior member

    Joined:
    Apr 4, 2012
    Messages:
    232
    Likes Received:
    3
    I did see F@H but it shouldn't cause problems. Try disabling it and then reboot. Also could you post your /var/log/Xorg.0.log.old please. It may provide some clues to your dilemma I would suggest you also post the suggested above my post.
     
  25. Red Squirrel

    Red Squirrel Lifer

    Joined:
    May 24, 2003
    Messages:
    36,925
    Likes Received:
    632
Loading...