(oops, I accidentally responded to you personally only. The emails are
supposed to go onto the list. I added the thread back to the list)

But is the config so big that memory usage is a concern here?

Also note, that the stuff that runs in main() is just generating a
streaming execution plan, which will be sent to the server. If the config
is really large, you might face timeouts during job submission.


On Fri, Jul 3, 2020 at 5:17 PM Jaswin Shah <[email protected]> wrote:

> It just results in more memory usage since, configs fetched by each flink
> job and they are going to store them in memory.
> ------------------------------
> *From:* Robert Metzger <[email protected]>
> *Sent:* 03 July 2020 20:31
> *To:* Jaswin Shah <[email protected]>
> *Subject:* Re: Custom service configs in flink
>
> Hi Jaswin,
>
> Usually, you have one Flink job per main() method (at least in our
> examples). However, you can use one (Stream)ExecutionEnvironment to submit
> multiple streaming jobs.
>
> Basically, the structure of your multi-job class would be
> public static void main(args) {
>  // 1. do rest call(s) to get business configs
> // 2. create execution environment
> // 3. assemble job topology
>  // 4. submit topology with env.execute() call. This special call clears
> the existing transformations, so that you can create another job in the
> same environment.
> env.execute(env.getStreamGraph("my job", true));
> // 5. assemble next job topology
> // ... repeat ...
> }
>
> However, I'm not sure if this approach isn't making things more
> complicated than necessary. Maybe you should just extract the REST-call
> logic into some separate class, that you are calling in each of your jobs?
>
> Best,
> Robert
>
> On Fri, Jul 3, 2020 at 12:06 PM Jaswin Shah <[email protected]>
> wrote:
>
> I have multiple flink jobs and have custom business configs which are
> shared between the job. Is it possible if one flink job loads configs in
> memory and all the flink jobs share the same configs? Basically, I am
> thinking to fetch configs in one flink job in memory via rest call which is
> one time and share those with all the jobs if possible. With this, I want
> to have an ability to update configs dynamically via kafka.
>
>

Reply via email to