[DRBD-user] single-disk nodes replicated in a multi-disk nas

Antonio Anselmi tony.anselmi at gmail.com
Thu Nov 18 18:50:35 CET 2010


I'm wondering if could be possible run DRBD in such scenario:

- n single disk nodes
- one nas equipped with n+1 disks
- all the disks have the same dimensions and geometry

so that in a 3-1 deployment (3 nodes and one nas with 4 disks) the
drbd.conf files could be the following

*node1 /etc/drbd.conf:
global { usage-count yes; }
common { syncer { rate 10M; } }
resource rA {
    protocol C;
    meta-disk internal;

    on node1 {
        address 10.0.0.1:7789;
        device /dev/drbd0
        disk /dev/sda;
    }
    on nas {
        address 10.0.0.10:7789;
        device /dev/drbd0
        disk /dev/sdb;
    }
}

*node2 /etc/drbd.conf:
global { usage-count yes; }
common { syncer { rate 10M; } }
resource rB {
    protocol C;
    meta-disk internal;

    on node2 {
        address 10.0.0.2:7789;
        device /dev/drbd0
        disk /dev/sda;
    }
    on nas {
        address 10.0.0.10:7789;
        device /dev/drbd1
        disk /dev/sdc;
    }
}

*node3 /etc/drbd.conf:
global { usage-count yes; }
common { syncer { rate 10M; } }
resource rC {
    protocol C;
    meta-disk internal;

    on node3 {
        address 10.0.0.3:7789;
        device /dev/drbd0
        disk /dev/sda;
    }
    on nas {
        address 10.0.0.10:7789;
        device /dev/drbd2
        disk /dev/sdd;
    }
}

*nas /etc/drbd.conf:
global { usage-count yes; }
common { syncer { rate 10M; } }

resource rA {
    protocol C;
    meta-disk internal;

    on node1 {
        address 10.0.0.1:7789;
        device /dev/drbd0
        disk /dev/sda;
    }
    on nas {
        address 10.0.0.10:7789;
        device /dev/drbd0
        disk /dev/sdb;
    }
}

resource rB {
    protocol C;
    meta-disk internal;

    on node2 {
        address 10.0.0.2:7789;
        device /dev/drbd0
        disk /dev/sdc;
    }
    on nas {
        address 10.0.0.10:7789;
        device /dev/drbd1
        disk /dev/sdd;
    }
}

resource rC {
    protocol C;
    meta-disk internal;

    on node3 {
        address 10.0.0.3:7789;
        device /dev/drbd0
        disk /dev/sda;
    }
    on nas {
        address 10.0.0.10:7789;
        device /dev/drbd2
        disk /dev/sdc;
    }
}

The single resources rA rB and rC are respectively defined in the
nodes node1, node2 and node3 and are grouped inside the drbd.conf of
the nas.

Is that correct?

Antonio


More information about the drbd-user mailing list