[DRBD-user] drbd master to slave synchronisation under heartbeat

avn avn at avn.ro
Mon Mar 8 18:22:31 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.

I tried to find a solution to my problem, but cannot find one either in the
forum or google.
I have 2 sets of 2 boxes each set in HA setup. One set is heartbeat 2.14
with drbd 8.2.7, the other heartbeat 3.0.2-rc2, pacemaker 1.0.7 and drbd
8.3.7. Both sets are gentoo 2.6.26 each box with drbd on top of raid1, box
in sets interconnected by exclusive use gigabit nic. Both sets have the same
config files:
global {
    usage-count no;
common {
  syncer { rate 100M; }
resource r0 {
  protocol C;
  handlers {
    pri-on-incon-degr "echo o > /proc/sysrq-trigger ; reboot -f";
    pri-lost-after-sb "echo o > /proc/sysrq-trigger ; reboot -f";
    local-io-error "echo o > /proc/sysrq-trigger ; reboot -f";
    outdate-peer "/usr/lib/heartbeat/drbd-peer-outdater -t 5";
  startup {
    degr-wfc-timeout 120;    # 2 minutes.
  disk {
    on-io-error   pass_on;
     fencing resource-only;
  net {
     sndbuf-size 512k;
     timeout       60;    #  6 seconds  (unit = 0.1 seconds)
     connect-int   10;    # 10 seconds  (unit = 1 second)
     ping-int      10;    # 10 seconds  (unit = 1 second)
     ping-timeout   80;    # 500 ms (unit = 0.1 seconds)
     max-buffers     2048;
     max-epoch-size  2048;
     ko-count 4;
     cram-hmac-alg "md5";
     shared-secret "somesecret";
     after-sb-0pri discard-younger-primary;
     after-sb-1pri violently-as0p;
     after-sb-2pri violently-as0p;
     rr-conflict call-pri-lost;
  syncer {
    rate 100M;
    al-extents 257;
  on lbl1 {
    device     /dev/drbd0;
    disk       /dev/md3;
    meta-disk  internal;
  on lbl2 {
    device    /dev/drbd0;
    disk      /dev/md3;
    meta-disk internal;

Now when something happened, like me doing a wrong config of a service in
heartbeat which makes the slave becoming master, or when master is powered
of unclean I bump into the following.
Some files in the new master are old versions (some time 1 month old), not
necessary in order of modifications (modified file 1 on 1st day, 2nd on
second, etc, 1 and 3 and 4 and 6 are ok, 2 and 5 old version). Before
plugging off the master I verified that status was Primary/secondary and
Uptodate aon master and Sec/primary also uptodate on slave. Nothing in logs.
This happen on both sets.
So I think I have a mistake somewhere, or I am not using the drbd right.

Any ideea?

Andrei Neagu
AVN Management
Bucharest, Romania
View this message in context: http://old.nabble.com/drbd-master-to-slave-synchronisation-under-heartbeat-tp27824570p27824570.html
Sent from the DRBD - User mailing list archive at Nabble.com.

More information about the drbd-user mailing list