nifi-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Matt Burgess <mattyb...@apache.org>
Subject Re: Processor disabled state not maintained across template download/import
Date Mon, 07 Nov 2016 16:10:15 GMT
I don't see DISABLED as being a runtime state, and I think it should
be included in the template export as it reflects a deliberate choice
by the author to prevent the processor from running when a flow is
started. It seems analogous to intentionally removing a required
property before exporting a template, to ensure that the processor
will be invalid on template load, and ensure that the user must
perform some action (enabling the processor or entering a required
property, respectively) before that processor can be started. Chris's
use case is IMO a good example of why the DISABLED state should be
retained.

Regards,
Matt

On Mon, Nov 7, 2016 at 10:34 AM, Oleg Zhurakousky
<ozhurakousky@hortonworks.com> wrote:
> This is actually a very interesting discussion to be had. . .
> So, at this point I believe that similar to other states of the component (i.e., RUNNING),
the DISABLED state is what is constituted as runtime-state and therefore should NOT be stored
into NOR expected to be restored from the template. Template IMHO should only be used to store
flow definitions.
> That said, I do see how someone may also treat/expect that DISABLED state is a special
state and as such does not represent the runtime state, rather the intention of the author
at the time of template export.
>
> So let’s duke it out here and summarize it in JIRA before we decide if any action needs
to be taken.
>
> Cheers
> Oleg
>> On Nov 7, 2016, at 10:20 AM, Joe Witt <joe.witt@gmail.com> wrote:
>>
>> Chris,
>>
>> I can see why we're not automatically starting processors when they're
>> placed on the graph but I do share your view that disabled processor status
>> should be honored and retained.  I think a JIRA for this is reasonable and
>> at the very least will get some good discussion and/or doc changes.
>>
>> Thanks
>> JOe
>>
>> On Mon, Nov 7, 2016 at 10:12 AM, McDermott, Chris Kevin (MSDU -
>> STaTS/StorefrontRemote) <chris.mcdermott@hpe.com> wrote:
>>
>>> If I create a template from a flow that has some disabled components, when
>>> I download and import that template into a different NiFi instance, the
>>> disabled state of those components is lost (they are no longer disabled.)
>>> I’m not sure when this information is being lost (is it saved in the
>>> template?)
>>>
>>>
>>>
>>> This makes using a template for deployment somewhat difficult.  Unless I’m
>>> missing something I am planning of entering a JIRA, but I wanted to check
>>> with the community first in case I am missing something.
>>>
>>>
>>>
>>>
>>>
>>> Thanks,
>>>
>>>
>>>
>>> Chris McDermott
>>>
>>>
>>>
>>> Remote Business Analytics
>>>
>>> STaTS/StoreFront Remote
>>>
>>> HPE Storage
>>>
>>> Hewlett Packard Enterprise
>>>
>>> Mobile: +1 978-697-5315
>>>
>>>
>>>
>>>
>

Mime
View raw message