[DRBD-user] Help: Replacing A Dead Node

Mike Sweetser - Adhost mikesw at adhost.com
Mon Jul 28 21:42:01 CEST 2008

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


I have a two-server DRBD primary/secondary configuration, running
Heartbeat.  Recently, the primary server had a drive failure and had to
be rebuilt, leaving the secondary server running the cluster (and quite
well, I may add).

The time has come to bring the original server back into the fold, which
means bringing it back into the DRBD fold, but WITHOUT damaging the
existing data.  This is my plan to do so - am I on the right track?  Am
I offbase?  Is this the right procedure?

* Copy /etc/drbd.conf from the working node to the new node
* Create the metadata on the new node with: drbd create-md all
* Attach the DRBD device on the new node: drbdadm attach all
* Make sure that the new node is in secondary mode: drbdadm secondary
all
* Connect the new node to the working node, discarding the existing data
so that it's overwritten by the good data: drbdadm -- --discard-my-data
connect all

Is this correct?  Will I have to take down Heartbeat while I reconnect?

Thanks for your help!

Mike Sweetser



More information about the drbd-user mailing list