Russ,
Given that the Content & FlowFile Repositories are intended to be rather short-lived, I'm not sure that you'd really have need to backup the repository (if you were to pause the flow in order to back it up, for instance, you could likely process the data just as fast as you could back it up - and at that point you'd be finished with it). That being said, for a production use case, I would certainly recommend running on a RAID configuration that provides redundancy so that if a disk were to go bad you'd still be able to access the data. For the Provenance Repository, there actually exists a Reporting Task that can send the data via Site-to-Site so that it can be exfilled however you see fit in your flow. Does this help? Thanks -Mark ________________________________ From: Russell Bateman <[email protected]> Sent: Friday, April 28, 2017 10:36 AM To: [email protected] Subject: Best practices for backing up an instance of NiFi? Been Googling hard on this (NiFi docs, Horton docs, these forums, stackoverflow) and I'm not seeing anything about specific wisdom surrounding backing up content_repository and flowfile_repository, why I would, why I wouldn't, turning NiFi off when I do, etc. I completely get the rest of the picture and have experience reproducing a NiFi instance, including flows, making sure that flows match data already in repositories, configuration, logs, etc. But, I was hoping to find wisdom on the rather huge undertaking of backing up the 4 repositories, must we pause NiFi when we do it, etc. I'm being asked for guidance by my IT staff who are ready to hop on this question now. I've been able to give them very pointed details on everything except the repositories. Any sharing is welcome. Thanks, Russ
