Hi,

There is no way to ensure that the primary node will always be a given
node. In case of failover or in case of restart, the primary node will
likely be a different node. That's why the validation is occurring on all
the nodes and that's why we recommend that all NiFi nodes are the same.

Hope this helps,
Pierre

2017-06-20 6:39 GMT+02:00 진유리 <[email protected]>:

> I have clustered NiFi v1.0.0 on 2 nodes.
>
> I added 'GetFile' processor and set scheduling strategy to 'On primary
> node'.
> On primary node, input directory exists but on another node, input
> directory does not exist.
> So I got error message like "'node2 : Input Directory' validated against
> '/var/test/' is invalid because Directory does not exist"
>
> I guess 'GetFile' processor check both node regardless of the scheduling
> strategy.
> Does it make a sense?
>
> I think it is not necessary to check all node when I set scheduling
> strategy to 'On primary node'.
>
>
> Thanks
> -Yuri Jin-
>

Reply via email to