[ https://issues.apache.org/jira/browse/HIVE-16484?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16317318#comment-16317318 ]
Marcelo Vanzin commented on HIVE-16484: --------------------------------------- Just amending my previous response: if the concern with {{SparkLauncher}} is the number of file descriptors because of the extra one used by the launcher server connection, using {{InProcessLauncher}} will probably end up decreasing the number of fds being used. Instead of potentially 3 fds for a child process (pipes for stdin / stdout / stderr), you have one for the socket connection. For the normal, child process case, then yes, you just get one extra file descriptor. (Or maybe you get even, because I think {{SparkLauncher}} will merge stdout and stderr in that case.) As for why this is better, I think the main advantage will come by using {{InProcessLauncher}} eventually, since Hive wouldn't need a separate Spark installation to be able to launch Spark apps. It could ship with everything ready to run HoS out of the box. Security can probably become simpler; instead of having to run kinit before starting a Spark child process, HS2 could potentially just instantiate {{InProcessLauncher}} inside a {{proxyUser.doAs}} call. I haven't actually tried that, but that's the general idea of how to use it in a secure env. > Investigate SparkLauncher for HoS as alternative to bin/spark-submit > -------------------------------------------------------------------- > > Key: HIVE-16484 > URL: https://issues.apache.org/jira/browse/HIVE-16484 > Project: Hive > Issue Type: Bug > Components: Spark > Reporter: Sahil Takiar > Assignee: Sahil Takiar > Attachments: HIVE-16484.1.patch, HIVE-16484.10.patch, > HIVE-16484.2.patch, HIVE-16484.3.patch, HIVE-16484.4.patch, > HIVE-16484.5.patch, HIVE-16484.6.patch, HIVE-16484.7.patch, > HIVE-16484.8.patch, HIVE-16484.9.patch > > > The {{SparkClientImpl#startDriver}} currently looks for the {{SPARK_HOME}} > directory and invokes the {{bin/spark-submit}} script, which spawns a > separate process to run the Spark application. > {{SparkLauncher}} was added in SPARK-4924 and is a programatic way to launch > Spark applications. > I see a few advantages: > * No need to spawn a separate process to launch a HoS --> lower startup time > * Simplifies the code in {{SparkClientImpl}} --> easier to debug > * {{SparkLauncher#startApplication}} returns a {{SparkAppHandle}} which > contains some useful utilities for querying the state of the Spark job > ** It also allows the launcher to specify a list of job listeners -- This message was sent by Atlassian JIRA (v6.4.14#64029)