This is a Jupyter based environment where we would like to put off binding a Spark session/context to the notebook until needed. In a YARN cluster simply bootstrapping the Spark context/session will require a couple of containers to be allocated which is wasteful unless the user really does perform (optional) Spark processing.
I opened a JIRA https://issues.apache.org/jira/browse/SPARK-20440 and attached PR 17731 to it as I think it better conveys both the problem and solution. Regards, Vin. On Sat, Apr 22, 2017 at 1:39 PM, Felix Cheung <felixcheun...@hotmail.com> wrote: > This seems some what unique. Most notebook environment, that I know of, > has a preset processing engine tied to the notebook; in other words when > Spark is selected as the engine then it is always initialized, not lazily > as you describe. > > What is this notebook platform you use? > > _____________________________ > From: Vin J <winjos...@gmail.com> > Sent: Saturday, April 22, 2017 12:33 AM > Subject: Re: [SparkR] - options around setting up SparkSession / > SparkContext > To: Felix Cheung <felixcheun...@hotmail.com> > Cc: <dev@spark.apache.org> > > > > This is for a notebook env that has the spark session/context bootstrapped > for the user. There are settings that are user specific so not all of those > can go into the spark-defaults.conf - such settings need to be dynamically > applied when creating the session/context. > > In Scala/Python, I would bootstrap a "spark" handle similar to what > spark-shell / psyspark-shell startup scripts do. In my case the > bootstrapped object could be of a wrapper class that took care of whatever > customization I needed while exposing the regular SparkSession > scala/python API. The user uses this object as he/she would use a regular > SparkSession to submit work to the Spark cluster. Since I am certain there > is no other way for users to perform Spark work except to go via the > bootstrapped object, I can achieve my objective of delaying creation of > SparkSession/Context until a call comes to my custom spark object. > > If I want to do the same in R, and let users write SparkR code as they > normally would, but bootstrapping a SparkContext/Session for them, then I > hit the issues as I explained earlier. There is no single entry point for > SparkContext/Session in SparkR API and so to achieve lazy creation of > SparkContext/session, it looks like the only option is to do some trickery > with the SparkR:::.sparkREnv$.sparkRjsc and SparkR:::.sparkREnv$.sparkRsession > vars. > > Regards, > Vin. > > On Sat, Apr 22, 2017 at 3:33 AM, Felix Cheung <felixcheun...@hotmail.com> > wrote: > >> How would you handle this in Scala? >> >> If you are adding a wrapper func like getSparkSession for Scala, and have >> your users call it, can't you do that same in SparkR? After all, while true >> you don't need a SparkSession object to call the R API, someone still needs >> to call sparkR.session() to initial the current session? >> >> Also what Spark environment you want to customize? >> >> Can these be set in environment variables or via spark-defaults.conf >> spark.apache.org/docs/latest/configuration.html#dynamically-loading- >> spark-properties >> >> >> _____________________________ >> From: Vin J <winjos...@gmail.com> >> Sent: Friday, April 21, 2017 2:22 PM >> Subject: [SparkR] - options around setting up SparkSession / SparkContext >> To: <dev@spark.apache.org> >> >> >> >> >> I need to make an R environment available where the >> SparkSession/SparkContext needs to be setup a specific way. The user simply >> accesses this environment and executes his/her code. If the user code does >> not access any Spark functions, I do not want to create a SparkContext >> unnecessarily. >> >> In Scala/Python environments, the user can't access spark without first >> referencing SparkContext / SparkSession classes. So the above (lazy and/or >> custom SparkSession/Context creation) is easily met by offering >> sparkContext/sparkSession handles to the user that are either wrappers on >> Spark's classes or have lazy evaluation semantics. This way only when the >> user accesses these handles to sparkContext/Session will the >> SparkSession/Context actually get set up without the user needing to know >> all the details about initing the SparkContext/Session. >> >> However, achieving the same doesn't appear to be so straightforward in R. >> From what I see, executing sparkR.session(...) sets up private variables in >> SparkR:::.sparkREnv (.sparkRjsc , .sparkRsession). The way SparkR api >> works, a user doesn't need a handle to the spark session as such. Executing >> functions like so: "df <- as.DataFrame(..)" implicitly access the private >> vars in SparkR:::.sparkREnv to get access to the sparkContext etc that are >> expected to have been created by a prior call to >> sparkR.session()/sparkR.init() etc. >> >> Therefore, to inject any custom/lazy behavior into this I don't see a way >> except through having my code (that sits outside of Spark) apply a >> delayedAssign() or a makeActiveBinding( ) on SparkR:::.sparkRsession / >> .sparkRjsc variables. This way when spark code internally references them, >> my wrapper/lazy code gets executed to do whatever I need done. >> >> However, I am seeing some limitations of applying even this approach to >> SparkR - it will not work unless some minor changes are made in the SparkR >> code. But, before I opened a PR that would do these changes in SparkR I >> wanted to check if there was a better way to achieve this? I am far less >> than an R expert, and could be missing something here. >> >> If you'd rather see this in a JIRA and a PR, let me know and I'll go >> ahead and open one. >> >> Regards, >> Vin. >> >> >> >> >> > > >
