[DRBD-user] flashcache + drbd + LVM2 + GFS2 + KVM live migration -> data corruption

Maurits van de Lande M.vandeLande at VDL-Fittings.com
Fri Feb 17 10:36:16 CET 2012

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


Hello,

I would like to use flashcache te speed up data reads from the drbd disk. Therefore  I have created a "write through" flashcache device
Using :
#/sbin/flashcache_create -p thru -b 16k cachedev /dev/sda1 /dev/sdc1

Then "/dev/mapper/cachedev" is used as a drbd target

...
on vmhost2a {
    device     /dev/drbd0;
    disk       /dev/mapper/cachedev;
    address    192.168.100.3:7788;
    meta-disk  internal;
  }
on vmhost2b {
    device    /dev/drbd0;
    disk      /dev/mapper/cachedev;
    address   192.168.100.4:7788;
    meta-disk internal;
}
...

As a test I have a VM which is being ping-ponged between two servers using live migration.
When I perform the test I get a corrupted VM!
See the following post

http://elrepo.org/bugs/view.php?id=228

Without flashcache it works fine.

Does anyone have a  working drbd+flashcache setup?
What do I do wrong?

Best regards,

Maurits van de Lande


| Van de Lande BV. | Lissenveld 1 | 4941VK | Raamsdonksveer | the Netherlands |T +31 (0) 162 516000 | F +31 (0) 162 521417 | www.vdl-fittings.com<http://www.vdl-fittings.com> |

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20120217/46a260ad/attachment.htm>


More information about the drbd-user mailing list