kvm hypervisor kernel panic

Question

My kvm hypervisor had been running with 20+ vps guests for a few months. But a few days ago it went to kernel panic. There were no abnormal operation that I found would cause the problem, except that the hypervisor was deleting a guest and shrubbing the corresponding logical volume at the time. Is this a hypervisor bug? How to fix it?

Apr 12 05:33:51 vps kernel: BUG: Bad page map in process crond  pte:00000f00     pmd:436a14058
Apr 12 05:33:51 vps kernel: addr:00008cbbca104512 vm_flags:08514270 anon_vma:(null) mapping:ffff41521ab409e0 index:16a
Apr 12 05:33:51 vps kernel: vma->vm_ops->fault: filemap_fault+0x0/0x500
Apr 12 05:33:51 vps kernel: vma->vm_file->f_op->mmap: ext4_file_mmap+0x0/0x60 [ext4]
Apr 12 05:33:51 vps kernel: Pid: 22373, comm: crond Not tainted 2.6.32-358.18.1.el6.x86_64 #1
Apr 12 05:33:51 vps kernel: Call Trace:
Apr 12 05:33:51 vps kernel: [<ffffffff8523ef18>] ? print_bad_pte+0x1d8/0x290
Apr 12 05:33:51 vps kernel: [<ffffffff8521b970>] ? generic_file_aio_read+0x380/0x700
Apr 12 05:33:51 vps kernel: [<ffffffff8523f03b>] ? vm_normal_page+0x6b/0x70
Apr 12 05:33:51 vps kernel: [<ffffffff8524179f>] ? unmap_vmas+0x61f/0xc30
Apr 12 05:33:51 vps kernel: [<ffffffff852476d7>] ?Apr 12 05:44:03 vps kernel: imklog 5.8.10, log source = /proc/kmsg started.

Memory failure? What does SEL (System Event Log) says? – SaveTheRbtz Apr 17 '14 at 06:54 — SaveTheRbtz, Apr 17 '14 at 06:54
'dmidecode -t system' shows that no error is detected – Purres Apr 17 '14 at 22:09 — Purres, Apr 17 '14 at 22:09

score 0 · Answer 1 · answered Apr 16 '14 at 21:28

It's a bug:

Apr 12 05:33:51 vps kernel: BUG: Bad page map in process crond  pte:00000f00     pmd:436a14058

You're running RHEL:

Apr 12 05:33:51 vps kernel: Pid: 22373, comm: crond Not tainted 2.6.32-358.18.1.el6.x86_64 #1

First, open a ticket with RH. If that doesn't bear fruit, update the kernel and pray it doesn't happen again. There's no "magic fix" to prevent this in the future.

kvm hypervisor kernel panic

1 Answers1