I am currently working out a design, where TN1 and TN2 Instances are synced. Same Data / Same Apps.
UseCase 1.) Disaster Recovery\Failover: TN1 fails → TN2 can startup all Apps and work with the latest synced data from replication task
UseCase 2.) Manually load balance the Apps and replicate to its peer TN Instance
Its not automatic failover and no failover in TrueNas HA/FA sense but an idea instead of having a 2nd instance of TN just for backup seems wasted Resources.
So this means, that App1 is writing Data on TN1 and App2 is writing data to TN2.
From practical way i can think of:
- all Apps need to be installed and configured with the same dataset layout on both TN instance
- Dataset replication will be configured to run from Active → Passive/DR Peer
- manual reverse replication (just create a replication task which is not sheduled )
- App Pool and Data Pool will have the same name/vdevs/mountpoints on both TN Instance
- for Apps which require .env i’ll consider to migrate to custom apps
So far so good, i’ll start less complex Apps and will make my way to more complex ones.
Any thoughst ideas .. objections ?