[DRBD-user] Primary / Unknown

[Felix Frank]

Hmm, in other words, you are purposefully putting your nodes in a split
brain situation?

It's a good thing you think of this case. Because by the nature of a
Primary/Primary setup, there can be no (safe) automatic split brain
recovery. As soon as the link drops, the datasets on your notes can
deviate rapidly.

You can look at
http://www.drbd.org/users-guide/s-configure-split-brain-behavior.html#s-automatic-split-brain-recovery-configuration
but there is really no sane policy to revocer from this situation.

This all is why you should never run a Primary/Primary setup without a
good STONITH implementation to protect you from data loss. Because
eventually you *will* have to discard the changes from one of the nodes.

I advise to carefully rethink the need for dual-primary operation.

HTH,
Felix

On 08/21/2013 09:56 PM, Victor Belizário wrote:
> Now, i'm in the test with the network. I did a "ifdown eth0" in the
> server01, and the cat /proc/drbd are like this :
> 
> version: 8.3.11 (api:88/proto:86-96)
> srcversion: DA5A13F16DE6553FC7CE9B2
> m:res   cs            ro               ds                 p  mounted  fstype
> 0:tray  WFConnection  Primary/Unknown  UpToDate/DUnknown  C
> 
> Ok, its waiting for connection, perfectly. So i turn on again the
> network, with ifup eth0 and think the state will be Primary / Primary
> again. But...not happens.