[DRBD-user] FAQ: Reconnecting after a temporary primary node failure

Felix Frank ff at mpexnet.de
Tue Mar 8 15:04:54 CET 2011


>       Oh. When you say reconnect do you mean reconnecting ethernet or
> drdbadm reconnect? If the former, too late and I will have to do it
> all again. If the latter, I have not done it yet so I should be good.
> In any case, here is the output with db1's ethernet cable reconnected:

No, you're right. I was referring to drbdadm connect.

> root at db1:~# drbdadm get-gi r0
> 3F197C90A31FC957:AD47F2C0FCB0CE1B:D8335248BF8F073F:201C5D8F032E15FB:1:1:1:0:0:0:0
> root at db1:~#
> 
> root at db2:~# drbdadm get-gi r0
> AD47F2C0FCB0CE1A:0000000000000000:D8335248BF8F073E:201C5D8F032E15FB:1:1:0:0:0:0:0
> root at db2:~#
> 
> So, from what you said, I *should* have
> 
> AD47F2C0FCB0CE1B
> AD47F2C0FCB0CE1A
> 
> the same, but as you can see the last character is different. Loads of fun.

That actually looks right. I'm don't clearly remember why the LSB is
always set on the Primary and never on the Secondary, but those actually
match.

Performing a connect on the Standalone node now should not result in
split-brain.

Cheers,
Felix


More information about the drbd-user mailing list