IBM Analytics Ideas

Welcome to the idea forum for IBM Analytics Clients! 

 

IBM Employees:

The correct URL for entering your ideas is https://hybridcloudunit-internal.ideas.aha.io

 

Clients:

Our team welcomes any feedback  and suggestions you have for improving our offerings / products!  This forum allows us to connect your offering / product improvement ideas with IBM product and engineering teams.

 

If you have not registered on this portal please click on the following link and register.  To complete registration you will need to open the email you will receive from Aha to confirm your identity.  https://ibm.biz/AnalyticsIdeasPortalRegister

 

YARN multitenant application separation and impersonation at YARN level

Current situation

  • A single Analytic Server (AS) instance can only create one spark context at YARN level.
  • Analytic Server does not support Impersonation at YARN level.

Bosch is planning to use the Analytic Server in a multiuser environment. In such an environment, different users create different streams that can have different priorities. One example for such priorities are “production” and “development”.  Streams that are needed in production (shop floor data) are more important (priority “production”) than streams that are created by developers that are developing new applications (priority “development”).

When a user first executes a modeler stream (submits a job), the analytic server generates a spark job. This spark job initializes a new YARN application, owned by the “as_user”. All subsequent jobs submits (by arbitrary users) are submitted into this previously started YARN application.

Consequently, SPARK jobs are not separated from each other. They are all part of one single YARN application. This can result in critical situations. If a single stream shows unwanted behavior it can compromise other streams. For example if a stream stalls and needs to be killed, the whole application needs to be killed. If the whole application fails, all streams in execution will fail. This is a very serious problem when running production streams next to development streams on the same cluster, as there is no separation of concerns.

In order to realize such a separation of concerns, the Analytic server provides multi tenancy concept. With this concept multiple Analytic Servers are started each AS creates its own YARN application. This concept, however, comes with the huge and unacceptable overhead of multiple Analytic Server instances.

Furthermore, Bosch also needs to provide SPSS to a wide range of different customers as a shared service. Therefore, Analytic Server must be capable of scheduling jobs from different customers into their specific resource pools. In this context, Bosch requires impersonation. In the best case, each stream that is submitted by a user is executed as an individual YARN application, owned by the respective user.

Bosch requirements to realize the use case described above:

  • Separate YARN (spark-)application for every job/stream which is being executed
    (separate application for every user, not tenant)
  • Impersonation of spark job at YARN level (not “as_user” but real “Bosch user”) as the “User” of the application (queue placement policy)
  • Definition of a specific resource pool (queues) for every tenant
    (Tenant <-> Queue mapping is necessary)
  • All functionality with only one Analytic Server instance per cluster
  • Guest
  • Jun 29 2018
  • Planned
Role Summary
  • Attach files
  • Steven Halter commented
    24 Jul 13:33

    The requirement:

    • Separate YARN (spark-)application for every job/stream which is being executed
      (separate application for every user, not tenant)

    is stated two ways in that sentence. A separate Yarn application for every stream and a separate Yarn application for every user. Which one is desired?

    For purposes of Spark caching, DataFrame reuse, it seems it would be better to have one Yarn application per user and route all jobs submitted by a single user to that application.

     

  • Guest commented
    08 Aug 08:07

    Every stream/job should run in a separate Yarn application. This way, every stream/job is isolated/separated by the other jobs and will not compromising them at resource level. In case a yarn-application will fail or needs to be killed because of a bug in the application/stream then only the single job within this application will be killed.

NOTICE TO EU RESIDENTS: per EU Data Protection Policy, if you wish to remove your personal information from the IBM ideas portal, please login to the ideas portal using your previously registered information then change your email to "anonymous@euprivacy.out" and first name to "anonymous" and last name to "anonymous". This will ensure that IBM will not send any emails to you about all idea submissions