On 3/11/20 1:58 PM, Mona Sinha wrote: > > There are two servers in cluster. One primary(as1) and other > secondary server(as2). Now if I try to follow this sequence: > > stop secondary (as2)-> stop primary (as1)-> start secondary > (as2)(expectation that it will get promoted to primary)
You are stopping the secondary while you have an active primary, so the secondary will have data that is not up to date with the data on the primary, therefore it is automatically marked as outdated - if you disconnect by administrator action (as opposed to the connection being lost) - if resource level fencing is configured (which you have in your configuration file) > It was found that there were some issues and the secondary server(as2) > could not be promoted as drbd stated that its state is not updated. ...which is correct and works as designed > Its state was either inconsistent/outdated. Outdated For a Secondary to be able to promote to Primary, one of the following must happen first: - Primary demotes to Secondary - Primary shuts down gracefully - Primary crashes - Replication link to the Primary is lost due to interrupted communication (with resource level fencing, this will still prevent the Secondary from promoting to Primary) br, Robert _______________________________________________ Star us on GITHUB: https://github.com/LINBIT drbd-user mailing list [email protected] https://lists.linbit.com/mailman/listinfo/drbd-user
