Hi Timothy,

Could you please elaborate more about your usecase? From my understanding,
it looks like you are executing some scripts as a one time run when there
is a request? You are not maintaining some "STATE".

If that's the case, I would suggest you to use the task framework of Helix.
The task will be executed when you submit to Helix. Regular state model
have "STATE" concept, it is a long run resource, more suited for a
distributed service.

Best,

Junkai


On Thu, Aug 27, 2020 at 6:25 PM Timothy Berger <[email protected]>
wrote:

> Hi Helix community,
>
> I've been getting familiarized with Helix with the intention of using it
> in a project I'm currently working on. We have been doing some
> prototyping with it using the Helix Agent to invoke a script that initiates
> a GRPC request to the locally running instance of our service. Something I
> didn't quite grok was, when the script has completed how is Helix made
> aware that the transition was successful or not? I didn't see in the agent
> source where the return code from the script was being used other than for
> logging purposes. Additionally while looking in the tutorials for the
> participant, an example transition handler looks like:
>
> *public void onBecomeOnlineFromOffline(Message message,
> NotificationContext context) { }*
>
> So from that signature there is no return value to indicate
> success/failure and there wasn't an example of how the handling function
> would signal back that a transition was successfully applied or not.
>
> So how would a participant signal back to the cluster that its state
> transition was successful/failed?
>
> Thank you for any help you may be able to give me to better understand.
> Very excited to use this.
>


-- 
Junkai Xue

Reply via email to