[GitHub] spark pull request #21996: [SPARK-24888][CORE] spark-submit --master spark:/...
Github user devaraj-kavali commented on a diff in the pull request:
https://github.com/apache/spark/pull/21996#discussion_r207763630
--- Diff: core/src/main/scala/org/apache/spark/deploy/SparkSubmit.scala ---
@@ -98,17 +98,24 @@ private[spark] class SparkSubmit extends Logging {
* Kill an existing submission using the REST protocol. Standalone and
Mesos cluster mode only.
*/
private def kill(args: SparkSubmitArguments): Unit = {
-new RestSubmissionClient(args.master)
- .killSubmission(args.submissionToKill)
+createRestSubmissionClient(args).killSubmission(args.submissionToKill)
}
/**
* Request the status of an existing submission using the REST protocol.
* Standalone and Mesos cluster mode only.
*/
private def requestStatus(args: SparkSubmitArguments): Unit = {
-new RestSubmissionClient(args.master)
- .requestSubmissionStatus(args.submissionToRequestStatusFor)
+
createRestSubmissionClient(args).requestSubmissionStatus(args.submissionToRequestStatusFor)
+ }
+
+ /**
+ * Creates RestSubmissionClient with overridden logInfo()
+ */
+ private def createRestSubmissionClient(args: SparkSubmitArguments):
RestSubmissionClient = {
+new RestSubmissionClient(args.master) {
+ override protected def logInfo(msg: => String): Unit =
printMessage(msg)
--- End diff --
I agree, user can change this log level. If the users configure the log
level as WARN or above(WARN is the default log level) then they can't see any
update/status from status and kill commands. I think we cannot expect the user
to configure the log level to INFO if they want to run status and kill commands
with status/update. Please let me know if you have any thoughts to fix this
better, I can make the changes. Thanks
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #21996: [SPARK-24888][CORE] spark-submit --master spark:/...
Github user felixcheung commented on a diff in the pull request:
https://github.com/apache/spark/pull/21996#discussion_r207717051
--- Diff: core/src/main/scala/org/apache/spark/deploy/SparkSubmit.scala ---
@@ -98,17 +98,24 @@ private[spark] class SparkSubmit extends Logging {
* Kill an existing submission using the REST protocol. Standalone and
Mesos cluster mode only.
*/
private def kill(args: SparkSubmitArguments): Unit = {
-new RestSubmissionClient(args.master)
- .killSubmission(args.submissionToKill)
+createRestSubmissionClient(args).killSubmission(args.submissionToKill)
}
/**
* Request the status of an existing submission using the REST protocol.
* Standalone and Mesos cluster mode only.
*/
private def requestStatus(args: SparkSubmitArguments): Unit = {
-new RestSubmissionClient(args.master)
- .requestSubmissionStatus(args.submissionToRequestStatusFor)
+
createRestSubmissionClient(args).requestSubmissionStatus(args.submissionToRequestStatusFor)
+ }
+
+ /**
+ * Creates RestSubmissionClient with overridden logInfo()
+ */
+ private def createRestSubmissionClient(args: SparkSubmitArguments):
RestSubmissionClient = {
+new RestSubmissionClient(args.master) {
+ override protected def logInfo(msg: => String): Unit =
printMessage(msg)
--- End diff --
this is not necessarily always the case - user can config log level easily?
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #21996: [SPARK-24888][CORE] spark-submit --master spark:/...
Github user devaraj-kavali commented on a diff in the pull request:
https://github.com/apache/spark/pull/21996#discussion_r207702588
--- Diff: core/src/main/scala/org/apache/spark/deploy/SparkSubmit.scala ---
@@ -98,17 +98,24 @@ private[spark] class SparkSubmit extends Logging {
* Kill an existing submission using the REST protocol. Standalone and
Mesos cluster mode only.
*/
private def kill(args: SparkSubmitArguments): Unit = {
-new RestSubmissionClient(args.master)
- .killSubmission(args.submissionToKill)
+createRestSubmissionClient(args).killSubmission(args.submissionToKill)
}
/**
* Request the status of an existing submission using the REST protocol.
* Standalone and Mesos cluster mode only.
*/
private def requestStatus(args: SparkSubmitArguments): Unit = {
-new RestSubmissionClient(args.master)
- .requestSubmissionStatus(args.submissionToRequestStatusFor)
+
createRestSubmissionClient(args).requestSubmissionStatus(args.submissionToRequestStatusFor)
+ }
+
+ /**
+ * Creates RestSubmissionClient with overridden logInfo()
+ */
+ private def createRestSubmissionClient(args: SparkSubmitArguments):
RestSubmissionClient = {
+new RestSubmissionClient(args.master) {
+ override protected def logInfo(msg: => String): Unit =
printMessage(msg)
--- End diff --
When `isInterpreter = true`(repl shell) the log is initialized and working
but the log level is setting to WARN due to that `RestSubmissionClient` logInfo
messages are not showing which are as part of the response. This PR change
effects when `isInterpreter = true` and for status/kill commands, and doesn't
change the other behavior.
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #21996: [SPARK-24888][CORE] spark-submit --master spark:/...
Github user felixcheung commented on a diff in the pull request:
https://github.com/apache/spark/pull/21996#discussion_r207702200
--- Diff: core/src/main/scala/org/apache/spark/deploy/SparkSubmit.scala ---
@@ -98,17 +98,24 @@ private[spark] class SparkSubmit extends Logging {
* Kill an existing submission using the REST protocol. Standalone and
Mesos cluster mode only.
*/
private def kill(args: SparkSubmitArguments): Unit = {
-new RestSubmissionClient(args.master)
- .killSubmission(args.submissionToKill)
+createRestSubmissionClient(args).killSubmission(args.submissionToKill)
}
/**
* Request the status of an existing submission using the REST protocol.
* Standalone and Mesos cluster mode only.
*/
private def requestStatus(args: SparkSubmitArguments): Unit = {
-new RestSubmissionClient(args.master)
- .requestSubmissionStatus(args.submissionToRequestStatusFor)
+
createRestSubmissionClient(args).requestSubmissionStatus(args.submissionToRequestStatusFor)
+ }
+
+ /**
+ * Creates RestSubmissionClient with overridden logInfo()
+ */
+ private def createRestSubmissionClient(args: SparkSubmitArguments):
RestSubmissionClient = {
+new RestSubmissionClient(args.master) {
+ override protected def logInfo(msg: => String): Unit =
printMessage(msg)
--- End diff --
doesn't this change the behavior even when the logger is
initialized/working?
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #21996: [SPARK-24888][CORE] spark-submit --master spark:/...
GitHub user devaraj-kavali opened a pull request:
https://github.com/apache/spark/pull/21996
[SPARK-24888][CORE] spark-submit --master spark://host:port --status
driver-id does not work
## What changes were proposed in this pull request?
In `SparkSubmit.scala` (`val uninitLog = initializeLogIfNecessary(true,
silent = true)`) -> `Logging.scala` (`val replLevel =
Option(replLogger.getLevel()).getOrElse(Level.WARN)`), the log level for
`rootLogger ` is overiri to `WARN ` but the status of the driver and kill
driver commands status have been logging with `INFO ` log level, so there is
nothing printing on the console for status and kill commands. This PR overrides
the `logInfo()` for `RestSubmissionClient` and redirects the msg to
`printStream`.
## How was this patch tested?
I verified it manually by running status and kill commands, these are the
results with and without the PR change.
- Without the PR change
```
[user1@user1-work-pc bin]$ ./spark-submit --master
spark://user1-work-pc:6066 --status driver-20180803165641-
[user1@user1-work-pc bin]$ ./spark-submit --master
spark://user1-work-pc:6066 --kill driver-20180803165641-
```
- With the PR change
```
[user1@user1-work-pc bin]$ ./spark-submit --master
spark://user1-work-pc:6066 --kill driver-20180803165641-
Submitting a request to kill submission driver-20180803165641- in
spark://user1-work-pc:6066.
Server responded with KillSubmissionResponse:
{
"action" : "KillSubmissionResponse",
"message" : "Driver driver-20180803165641- has already finished or
does not exist",
"serverSparkVersion" : "2.4.0-SNAPSHOT",
"submissionId" : "driver-20180803165641-",
"success" : false
}
[user1@user1-work-pc bin]$ ./spark-submit --master
spark://user1-work-pc:6066 --status driver-20180803165641-
Submitting a request for the status of submission
driver-20180803165641- in spark://user1-work-pc:6066.
Server responded with SubmissionStatusResponse:
{
"action" : "SubmissionStatusResponse",
"driverState" : "FINISHED",
"serverSparkVersion" : "2.4.0-SNAPSHOT",
"submissionId" : "driver-20180803165641-",
"success" : true,
"workerHostPort" : "xx.x.xx.xxx:42040",
"workerId" : "worker-20180803165615-10.3.66.149-42040"
}
```
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/devaraj-kavali/spark SPARK-24888
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/spark/pull/21996.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #21996
commit 1e8000ffacabac742d16efee72ec1e421225a272
Author: Devaraj K
Date: 2018-08-04T00:47:23Z
[SPARK-24888][CORE] spark-submit --master spark://host:port --status
driver-id does not work
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
