I am seeing very similar issues with 09.04 (and I believe I also had problems under 08.10).
In my case I see episodes of very high load average and intense disk thrashing, which usually lead to a functional lock-up of the system. I am unable to change virtual terminals, to move the mouse pointer, to kill X with ctrl-alt-bksp, or even to ssh in from another system (times out during the negotiation). The only solution is to force a semi- orderly shutdown using magic-sysrq key sequences. Once I kill all processes using magic-sysrq, the thrashing stops. When this occurs, on the odd occasion I have been able to kill X, sanity is restored and the thrashing stops. I believe that one of my X apps is triggering this bad behaviour - most likely part of KDE 4.2, or perhaps the Intel X server itself. Both KDE4 and the Intel X server are seeing a lot of active development, so some regressions are to be expected. My hardware is an HP Pavilion 1680ea (Core Duo, 2GB RAM). I have no swap configured. (I have tried configuring swap, and it doesn't appear to solve my issue.) What could be causing this crazy disk I/O? A typical scenario is as follows - I am running a KDE4 session, machine is fairly freshly booted (10 mins uptime), I have Firefox open with one tab (no fancy content). Memory use, excluding cache, is 400MB used out of 2GB physical (as shown by htop). I open some other app like Thunderbird or Okular and the load average spikes to 10 or so and the disk starts thrashing. I have observed htop and iotop while this is going on. htop does not show memory exhaustion - memory use maybe peaks at 600MB or 800MB before everything becomes so unresponsive that htop no longer updates. iotop does not show excess disk i/o from any particular process. I have no swap configured, so the I/O can't be swapping per se. I don't believe it is a storm of application reads or writes (iotop would show this). Could it be the system frantically demand paging in executables / libs? Maybe memory got suddenly exhausted before htop could show it? (Can anything allocate 1.5GB in a second or so, I wonder?) I am not sure what information to provide really. Debugging this is driving me crazy (I've been trying for a while). I can't seem to instrument the problem in order to work out what is going on when it all comes unravelled. b...@gecko2:/var/log$ uname -a Linux gecko2 2.6.28-11-generic #42-Ubuntu SMP Fri Apr 17 01:57:59 UTC 2009 i686 GNU/Linux I have attached the output of lspci -vvxx ** Attachment added: "lspci -vvxx output" http://launchpadlibrarian.net/25866482/lspci.txt -- [Hardy] High system load coming out of suspend and hibernate https://bugs.launchpad.net/bugs/211153 You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
