Out of curiosity I tried an Arch Linux instance (running Linux 3.5.6-1-ARCH), which also has CONFIG_SCHED_AUTOGROUP:
# zgrep AUTOGROUP /proc/config.gz CONFIG_SCHED_AUTOGROUP=y I ran the same pgslam workload on it, and it filled 64G of the /var/lib/postgres md-raid before I stopped it. This indicates to me one of two possibilities. Either this scheduler deadlock has been fixed between 3.2.x and 3.5.x, or one of the Ubuntu-specific commits is to blame. Unfortunately it's not super-easy to bisect this problem because it requires a lot of kernel rebuilds and lots of patience. I'm working on some other things at the moment, but hopefully we can narrow this down further and get it fixed properly. Autogrouping is nice, and I'd hate to toss it out. Perhaps one way to narrow it down would be to try the same workload on a 12.10 Quantal instance. I'll try to kick that off in the next couple hours. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1011792 Title: Kernel lockup running 3.0.0 and 3.2.0 on multiple EC2 instance types To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1011792/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
