[DRBD-user] drbd invalidate and drbd-partitions

Nadezhda S. Zhivchikova ming at pereslavl.ru
Tue Feb 16 07:44:07 CET 2010

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


Hello.

I have a 2 node with kernel 2.6.20, DRBD 0.7.23, LVM 2.01.04-5.

I want to move Secondary node from one server (192.168.48.2) to another
(192.168.48.3).

I do:

on the new node:

lvcreate -L1000 -n test vg0
lvcreate -L128 -n test-drbd-meta vg0


on the Primary node (192.168.48.1) I change the configuration:

old configuration:

resource "test" {
  protocol C;
  incon-degr-cmd "/bin/false";
  syncer { rate 50M; }
  startup { degr-wfc-timeout 5; wfc-timeout 20; }
  on node1 {
    device     /dev/drbd/test;
    disk       /dev/vg0/test;
    address    192.168.48.1:7821;
    meta-disk  /dev/vg0/test-drbd-meta[0];
  }
  on node2 {
    device     /dev/drbd/test;
    disk       /dev/vg0/test;
    address    192.168.48.2:7821;
    meta-disk  /dev/vg0/test-drbd-meta[0];
  }
}

new configuration:

resource "test" {
  protocol C;
  incon-degr-cmd "/bin/false";
  syncer { rate 50M; }
  startup { degr-wfc-timeout 5; wfc-timeout 20; }
  on node1 {
    device     /dev/drbd/test;
    disk       /dev/vg0/test;
    address    192.168.48.1:7821;
    meta-disk  /dev/vg0/test-drbd-meta[0];
  }
  on node3 {
    device     /dev/drbd/test;
    disk       /dev/vg0/test;
    address    192.168.48.3:7821;
    meta-disk  /dev/vg0/test-drbd-meta[0];
  }
}

I copy drbd.conf on new node.

on old Secondary node:
drbdadm down test

on Primary node:
drbdadm adjust test

on new Secondary node:
drbdadm up test

I see in /proc/drbd after synchronization: 

 0: cs:Connected st:Primary/Secondary ld:Consistent  
    ns:1024000 nr:0 dw:3081256 dr:6169436 al:15 bm:973 lo:0 pe:0 ua:0
ap:0
 1: cs:Unconfigured
 2: cs:Unconfigured


Then I try to compare the partitions of the nodes. The partitions are
different:

node1:~# ssh user at node3 "dd if=/dev/vg0/test" | cmp - /dev/vg0/test
- /dev/vg0/test differ: char 65541, line 1
2048000+0 records in
2048000+0 records out
1048576000 bytes transferred in 25.108733 seconds (41761406 bytes/sec)

I say on new Secondary node:
drbdadm invalidate test

When Resync has finished, I try to compare the partitions on the nodes.
The partitions are different too.

After invalidate DRBD will copy each and every block over from its peer,
to bring the local storage device back in sync. Does it mean that the
partitions must be the same after invalidate?


Thanks for any help,
Nadezhda.




More information about the drbd-user mailing list