Joe, Sorry for the confusion, I meant to say, like a single network mounted drive accessible to all nodes and all nodes pointing to one “content_repository” directory in that mount. Can it be done?
-Karthik From: Joe Witt [mailto:[email protected]] Sent: Wednesday, December 20, 2017 4:49 PM To: [email protected] Subject: RE: [EXT] Re: NiFi Clustering karthik Terminology wise you cannot share a repository across nodes. So that is a big con :) But if you mean can i share something like a network mounted drive by giving each node its own mount pointa for its repos then sure that works fine. In such a case you need to have strong iops provisioning, you need to realize there will be higher roundtrip times for rw ops. But in general this can be managed. Thanks joe On Dec 20, 2017 6:45 PM, "Karthik Kothareddy (karthikk) [CONT - Type 2]" <[email protected]<mailto:[email protected]>> wrote: Bryan , Thanks for your response , We are going with individual repositories. Can you explain what are the cons of having a shared content repository between multiple nodes ? I may be asking a dumb question but we are trying to understand if shared filers increase the throughput of the system. Thanks Karthik -----Original Message----- From: Bryan Bende [mailto:[email protected]<mailto:[email protected]>] Sent: Wednesday, December 20, 2017 11:54 AM To: [email protected]<mailto:[email protected]> Subject: [EXT] Re: NiFi Clustering Hello, Each node must have its own repositories, you can't point multiple nodes at the same repositories. The repositories for a given node could be on local disk, or on a mounted network drive, but local disk will likely have the best performance. Thanks, Bryan On Wed, Dec 20, 2017 at 1:25 PM, Karthik Kothareddy (karthikk) [CONT - Type 2] <[email protected]<mailto:[email protected]>> wrote: > All, > > > > We are trying to setup a new NiFi (1.4.0) cluster with 3 nodes, and > trying to decide on how should we proceed with the repository > structure. Should we be pointing the same repositories for all nodes > (shared filers accessible by all 3 nodes) or give separate local > repositories for all three nodes and let each node has its own data > copy. Is there any recommendation for this for better performance and > throughput? > > > > Thanks for your time. > > > > -Karthik
