The message seems problem is on disk io. This "child_rip" handler erases everything in shutting down sequences, but kernel object already erased. Its cause from .. 1. some driver double freed a same object. (driver problem) 2. heavy new/delete events double frees a same object(environment problem) To resolve your problem, keep your disk access stably. I saw you use xen kernel to ask xen-expert. HOST and GEST can't handle like memory problem on VM environment, Reduce I/O, Change your hypervisor. you can try many version of kernel-xen on centos.(or Xenserver) you can reduce your traffic on your environment.(or upgrade your system ;) -Tsuyoshi