Joey,

On every connection and process now you can right-click and see stats.
Those stats reveal how often things are failing, etc.  Right now you'd not
be able to distinguish between a permissions issue as the cause and
something more transient.  But, we should put the failure reason as an
attribute of a flow file so one could route on that should the distinction
help them make a decision or better track.

Thanks
JOe

On Mon, Dec 22, 2014 at 1:21 PM, Joey Echeverria <[email protected]> wrote:

> Neat, I hadn't thought of using that as a retry mechanism. Is there any
> way to tell how often a file has cycled through failure so you could
> distinguish between a transient failure and a permanent failure, like a
> permissions issue?
>
>
> -Joey
>
>
> —
> Joey Echeverria
>
> On Mon, Dec 22, 2014 at 10:18 AM, Mark Payne <[email protected]> wrote:
>
> > Joey,
> > Nice - thanks for putting that together & sending out!
> > Only feedback I would give is that for the PutHDFS, in the typical use
> case you wouldn't want to Auto-terminate the failure relationship. Instead,
> you'd route the failure relationship back to the PutHDFS Processor. This
> way, if NiFi can't push the file to HDFS (network problem, typo in config
> file, whatever), NiFi will be able to retry until it's successful.
> > Thanks-Mark
> >> Date: Mon, 22 Dec 2014 10:06:28 -0800
> >> From: [email protected]
> >> To: [email protected]
> >> Subject: Blog post
> >>
> >> I wrote a quick getting started blog post on NiFi for the ingest.tips
> blog here:
> >>
> >> http://ingest.tips/2014/12/22/getting-started-with-apache-nifi/
> >>
> >>
> >> I focused on intefration with Hadoop and the flexibility of NiFi to
> integrate existing ETL flows and tools.
> >>
> >>
> >> -Joey
> >>
> >> —
> >> Joey Echeverria
> >
>

Reply via email to