Thank you.
Is there any documentation on the ideal cluster set up for production
environment? I saw the 3 zoo keepers on 3 VM's and 3 nodes on 3 VM's is the
best. Could you confirm?
Also the ideal system requirement for each node and each zoo keeper.....in
terms of CPU, memory, disk space etc?
and finally archiving....logs, flow files etc and their frequency

Thanks
Anil


On Tue, Mar 7, 2017 at 3:11 PM, Bryan Bende <[email protected]> wrote:

> Anil,
>
> The short answer is that each node in the cluster is running the
> entire data flow you see in UI, and NiFi does not transfer flow files
> between nodes unless the data flow is created to do that.
>
> This article describes approaches to distributing data across the
> cluster, which is dependent on how the data is entering the cluster:
> https://community.hortonworks.com/articles/16120/how-do-i-
> distribute-data-across-a-nifi-cluster.html
>
> Thanks,
>
> Bryan
>
>
> On Tue, Mar 7, 2017 at 3:03 PM, Anil Rai <[email protected]> wrote:
> > Thanks Andy. I am looking for the processing of a flow file.If i create a
> > flow, with multiple processor, how does this request flow in nifi in a
> > cluster.
> >
> >
> > On Tue, Mar 7, 2017 at 2:58 PM, Andy LoPresto <[email protected]>
> wrote:
> >
> >> Anil,
> >>
> >> Are you talking about processing a flowfile or responding to a REST HTTP
> >> request? The Admin Guide section on clustering [1] offers more
> information
> >> but if you can clarify your question you may get better answers from the
> >> mailing list.
> >>
> >> [1] https://nifi.apache.org/docs/nifi-docs/html/administration-
> >> guide.html#clustering
> >>
> >>
> >> Andy LoPresto
> >> [email protected]
> >> *[email protected] <[email protected]>*
> >> PGP Fingerprint: 70EC B3E5 98A6 5A3F D3C4  BACE 3C6E F65B 2F7D EF69
> >>
> >> On Mar 7, 2017, at 11:48 AM, Anil Rai <[email protected]> wrote:
> >>
> >> Team,
> >>
> >> Is there any documentation that explains how Nifi processes a request
> once
> >> it reach nifi in a cluster settings?
> >>
> >> Regards
> >> Anil
> >>
> >>
> >>
>

Reply via email to