On Wed, 2 Apr 2014, John R Pierce wrote: > I have a server thats been running fine for a year or two lock up a few > times recently, requiring power cycling. > > The /var/log/messages after a lockup last night is appended to this > message. > > hardware is a pretty typical server, Supermicro X8DTE-F motherboard, > dual Xeon X5650, 48GB ECC memory, LSI SAS 2008 for the boot disks, > and LSI MegaRAID SAS 9261-8i for the data volume. Lots of 3TB disks > in a raid60. Primary application is BackupPC v3.3.0 (from EPEL), it > also has an NFS export (also used for backup purposes). > > Runs CentOS 6.latest (kernel 2.6.32-431.11.2.el6.x86_64). X is not > loaded (inittab level 3). selinux is permissive, iptables is not > loaded. this server is on a corporate internal network, 1 Intel > 82574L NIC configured with static IP, 2nd one is not in use. The lovely Supermicro 82574L bug! There was a similar thread about a month ago: http://lists.centos.org/pipermail/centos/2014-March/141348.html The short answer that's worked for me: Add pcie_aspm=off to your boot-time kernel options. I've also started running the kmod-e1001e package from elrepo.org. Also, if MSI-X isn't already turned off in the BIOS, others have suggested making sure it is. -- Paul Heinlein heinlein at madboa.com 45°38' N, 122°6' W