[DRBD-user] Concurrent local write detected

Florian Haas florian at hastexo.com
Mon Jun 25 17:10:51 CEST 2012

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


On Fri, Jun 22, 2012 at 6:53 PM, Thilo Uttendorfer
<t.uttendorfer at linux-ag.com> wrote:
> Hi,
>
> we have a KVM cluster with each virtual machine running on top of a DRBD
> device. After running this setup for many month we recently saw these log
> messages on two (of total 15) DRBD devices:
>
> Jun 22 17:53:01 server-v1 kernel: [4995041.347749] block drbd13:
> qemu-system-x86[9395] Concurrent local write detected! [DISCARD L] new:
> 53981575s +28672; pending: 53981575s +20480

Those writes overlap only partially, so yes that's cause for concern.
Not saying that you are actively shredding your data, but you at least
need to double check.

> Looking at older posts on this topic [1], we are worried that we have serious
> data corruption. Does this indicate that qemu-system-x86 (kvm) has a bug in
> some (probably rare) conditions?

Libvirt configuration or qemu/kvm command line please? (don't post
them here; pastebin and share the URL instead).

Cheers,
Florian

-- 
Need help with High Availability?
http://www.hastexo.com/now



More information about the drbd-user mailing list