Re: [gyrex-dev] Gyrex Jobs

[Gunnar Wagenknecht <gunnar@xxxxxxxxxxxxxxx>]

Hi Jürgen,

> Am 17.12.2015 um 11:16 schrieb Jürgen Albert <j.albert@xxxxxxxxxxxxxxxxxx>:
> Another question in this context: How can I realize a job that should be executed on every node of a certain type? For example one node saves an object and every node that holds an index should run a job, that takes the Object out of the DB and updates the local index.

It sounds like this isn't a good use case for Gyrex distributed jobs. This sounds more like a use case for an distributed event bus. Every time an event is received, a local thread could perform the indexing.

As for distributed jobs, I would expect the implementation to be different - one job builds the index and then distributes it to all the nodes requiring the new index. The distribution could happen via a shared drive or using scp or some other protocol.

-Gunnar