#175 PR merged: Update usr/share/rear/layout/prepare/GNU/Linux/15_include_drbd_code.sh

dragon299 opened issue at 2012-11-06 15:17:

Please have a look on https://github.com/rear/rear/issues/174

jhoekx commented at 2012-11-07 09:10:

Gratien's observation seems correct to me.

And now I remember the reason why I used attach instead of up. Using up did not work when the slave is not online yet. Did you test that case? It looks to me like the only time you want to restore a master (primary) is when both nodes are down.

dragon299 commented at 2012-11-07 15:15:

When the slave is online, up also worked in my test enviroment, but I get a split brain.

dragon299 commented at 2012-11-08 09:30:

I tested it with two drbd nodes and one drbd device (active/passive) in a vmware environment. The restore works well in all test cases.

[Export of Github issue for rear/rear.]