[DRBD-user] data corruption scenario in ha cluster

Mia Lueng xiaozunvlg at gmail.com
Wed Feb 23 17:13:00 CET 2011

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


Hi:
   I am providing a HA solution base on hb+drbd8. And I have a
question how hb+drbd prevent data corruption when the following events
happen.

Suppose there are two nodes A, B in the cluster.

1. A is primary and B is slavery.   The state transforming like this:
A(running) B(running)---->A(running)
B(hangup)---->A(hangup)B(hangup)--->A(hangup)B(start)

When B starts, How can it know that the data will be corrupt and do
not start as primary DRBD.

2. A is primary and B is slavery. The drbd net is broken. Before B
know this( the drbd state switch to Inconsistent), A write some data
to drbd disk and then  hangup. Then how can B know that the data will
be corrupt and do not start as primary DRBD.

Best Regards
Mia Lueng



More information about the drbd-user mailing list