hadoop - Spring XD dynamic deployment manifest -
I'm reading the Spring XD documentation quite heavily and really do not want to catch up with two things I want Hadoop Yarn Received in relation to
It may be that they are not yet supported or will never be supported - probably because I am missing something which makes my scenarios non-sensational ... < / P>
- It is possible for yarn in the hoodop that application mas Specific hosts by Su Rack 'awareness should be requested to allocate containers. It allows to process data stored on HDFS.
Can such functionality be displayed as a valuated asset in a property?
Note that I'm not talking about sections with partition, where the same container holds the same message for all the modules in the stream.
There are several instances of a module in the center of the section posted on the container set. In these containers, there will also be segments of large static data already present in these containers. I would like to have all the allocation based on an evaluated rule The most appropriate module example should be for example, in which each message is being processed in a pre-existing large file stored on one of the containers.
- Is it possible to dynamically 'scale out' a module in more containers after deploying on the stream? For example, if a module is being interrupted once in the stream, then if the stream is deployed, then the number of examples of that module will be increased in more containers dynamically?
Thanks to Nick
. For example, if a module is being interrupted once in the stream, then once the stream can be deployed, then the number of examples of that module will be increased in more containers dynamically?
The direct binding of the module helps in your case allowing the direct binding section's modules to be co-located on the same container and each container will be able to move the co-located modules to dynamic deployment Provides which connects to the cluster. You can find more information about this
Comments
Post a Comment