[DRBD-user] Avoid split brain in a dual primary configuration with intelligent switches

Mario Giammarco mgiammarco at gmail.com
Thu Mar 19 12:12:47 CET 2009

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

Lars Ellenberg <lars.ellenberg at ...> writes:

> dual primary: both nodes potentially have "application writes 
in flight".
> replication link loss means, that at this point,
> in the general case, we have already diverging data sets.
> not one side is stricktly ahead/behind the other.
> but both sides differ in some non-trivial way.

Ok it is now clear to me the dual primary additional problem of 
dual nodes writing.

Now I suppose that with write quorum you will mark one of the two nodes
 dead and use only the other one. So you declare the "writes in flight"
 of the dead node failed and the upper layer must redo them. So can I
 simulate this behaviour with heartbeat and fencing (I am in a lucky case
because I use drbd as an iscsi target so the "upper layer" is the 
iscsi client on another server which  can do multipathing on the
 two nodes) ?

More information about the drbd-user mailing list