ClusterID getting out of sync with NameNode, StandBy NN and DataNode

[#1]

Running a cluster on Azure where twice now something has failed to start due to the ClusterID being out of sync. The first time it was the DataNodes failed to start as the ClusterID was different to the NN. Then, after enabling HA, the Standby NN and all DN's started, however the Active NN failed (which then caused active to failover to standby). After manually syncing the ID's the cluster started fine in both cases. We didn't manually format the NN either time so am seriously confused as to how this could have happened.

What possible ways can the Cluster ID be changed?

[#2 · Accepted solution]

I am fairly certain this was caused by the disks being remounted in different positions on restart. http://sebastiancarroll.github.io/2016/10/24/azure-for-hdp-deployments.html