Restore DR using only .kdb files

We have an issue with Appian's described backup and restore process.  Our VM backup/snapshot contains the server/kafka-logs directory.  Our shared disk (HA configuration) backup, however, is near realtime and so we would actually like to execute checkpoints at less busy times so that we always have a fresh (few hours old) set of .kdb files to which we can restore during a DR scenario.  However, since the VM Backup/snapshot is only once per day (restriction of the cloud we're on), the kafka-logs and the subsequent checkpoint .kdb files would become out of synch, with the .kdb files being newer.

In this scenario, will Appian figure out that the kafka logs are old and not use them?  Or will the lack of synchronization cause the system to fail?  Is there a way that we could "empty" the kafka-logs, so that only the .kdb files are used?

Thank you,


  Discussion posts and replies are publicly visible

Parents Reply
  • Hi Jeff,
    we have a similar configuration related to the location of <APPIAN_HOME>/data-server/data. We asked Appian through a support ticket to get confirmation:
    1. Is the content in data-server/data necessary to bring up the Platform? Answer: Yes (we expected this, data server is kdb)

    2. Has the directory <APPIAN_HOME>/data-server/data to be on a local disk? Answer: No

    So we are going to test moving data-server/data to NAS drive.

    Just to avoid any confusion: for the Kafka logs directory 'services/data/kafka-logs' and 'services/data/zookeeper' Appian
    recommends to have those residing on a local disk .