|Product Version||6.2|| |
|Target Version||Fixed in Version|| |
|Summary||0005729: Kernel panics and server freezes with WARNING: at arch/x86/kernel/smp.c:117 native_smp_send_reschedule...|
|Description||We have had this machine running CentOS with several KVMs for months. Several weeks ago we decided to add redundancy by creating a DRBD cluster. About two days after finishing the DRBD setup, we saw a kernel panic (the first image in the imgur album: http://imgur.com/a/MGAwx). As you can see in the image, we were running 2.6.32-220.4.1.el6.x86_64. After updating the kernel to 2.6.32-220.13.1.el6.x86_64, the server ran for about four days before completely freezing again (The second image in the album). And then again about 4 days later we saw the same panic (image three). It has been running for about a day and half since the last panic. |
I have searched all over online and asked around on IRC and I can't find anyone that has seen a message like this. Some suggested that it might have to do with the huge page splitting, since the first panic references huge_mem.c, however the second two are on a different kernel version and show errors in smp.c instead. I'm unsure if DRBD is part of the problem because it only started happening shortly after that setup. At this point I'm at a loss. I'd like to get this figured out ASAP since downtime is bad.
|Additional Information||The machine is a HP Proliant DL385 G7 with dual AMD Opertron 6128 processors and 32GB memory.|
|Tags||6.2, drbd, freeze, kernel, kernel panic, kvm, panic, smp|