[jira] [Commented] (SPARK-31482) spark.kubernetes.driver.podTemplateFile Configuration not used by the job
[
https://issues.apache.org/jira/browse/SPARK-31482?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17087195#comment-17087195
]
Ed Mitchell commented on SPARK-31482:
-
Yeah. It's actually there but there's some unescaped HTML that was fixed in a
later commit:
[https://github.com/apache/spark/commit/44e314edb4b86ca3a8622124539073397dbe68de#diff-b5527f236b253e0d9f5db5164bdb43e9]
> spark.kubernetes.driver.podTemplateFile Configuration not used by the job
> -
>
> Key: SPARK-31482
> URL: https://issues.apache.org/jira/browse/SPARK-31482
> Project: Spark
> Issue Type: Bug
> Components: Kubernetes
>Affects Versions: 3.0.0
>Reporter: Pradeep Misra
>Priority: Blocker
>
> Spark 3.0 - Running Spark Submit as below and point to a MinKube cluster
> {code:java}
> bin/spark-submit \
> --master k8s://https://192.168.99.102:8443 \
> --deploy-mode cluster \
> --name spark-pi \
> --class org.apache.spark.examples.SparkPi \
> --conf spark.kubernetes.driver.podTemplateFile=../driver_1E.template \
> --conf spark.kubernetes.executor.podTemplateFile=../executor.template \
> --conf spark.kubernetes.container.image=spark:spark3 \
> local:///opt/spark/examples/jars/spark-examples_2.12-3.0.0-preview2.jar 1
> {code}
>
> Spark Binaries - spark-3.0.0-preview2-bin-hadoop2.7.tgz
> Driver Template -
> {code:java}
> apiVersion: v1
> kind: Pod
> metadata:
> labels:
> spark-app-id: my-custom-id
> annotations:
> spark-driver-cpu: 1
> spark-driver-mem: 1
> spark-executor-cpu: 1
> spark-executor-mem: 1
> spark-executor-count: 1
> spec:
> schedulerName: spark-scheduler{code}
> Executor Template
>
> {code:java}
> apiVersion: v1
> kind: Pod
> metadata:
> labels:
> spark-app-id: my-custom-id
> spec:
> schedulerName: spark-scheduler{code}
> Kubernetes Pods Launched - Two Executor Pods were launched which was default
> {code:java}
> spark-pi-e608e7718f11cc69-driver 1/1 Running 0 10s
> spark-pi-e608e7718f11cc69-exec-1 1/1 Running 0 5s
> spark-pi-e608e7718f11cc69-exec-2 1/1 Running 0 5s{code}
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-31482) spark.kubernetes.driver.podTemplateFile Configuration not used by the job
[
https://issues.apache.org/jira/browse/SPARK-31482?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17087167#comment-17087167
]
Pradeep Misra commented on SPARK-31482:
---
Have created another bug for documentation -
https://issues.apache.org/jira/browse/SPARK-31481
As I have understood, pod template work was done as part of this commit -
[https://github.com/apache/spark/commit/f6cc354d83]
> spark.kubernetes.driver.podTemplateFile Configuration not used by the job
> -
>
> Key: SPARK-31482
> URL: https://issues.apache.org/jira/browse/SPARK-31482
> Project: Spark
> Issue Type: Bug
> Components: Kubernetes
>Affects Versions: 3.0.0
>Reporter: Pradeep Misra
>Priority: Blocker
>
> Spark 3.0 - Running Spark Submit as below and point to a MinKube cluster
> {code:java}
> bin/spark-submit \
> --master k8s://https://192.168.99.102:8443 \
> --deploy-mode cluster \
> --name spark-pi \
> --class org.apache.spark.examples.SparkPi \
> --conf spark.kubernetes.driver.podTemplateFile=../driver_1E.template \
> --conf spark.kubernetes.executor.podTemplateFile=../executor.template \
> --conf spark.kubernetes.container.image=spark:spark3 \
> local:///opt/spark/examples/jars/spark-examples_2.12-3.0.0-preview2.jar 1
> {code}
>
> Spark Binaries - spark-3.0.0-preview2-bin-hadoop2.7.tgz
> Driver Template -
> {code:java}
> apiVersion: v1
> kind: Pod
> metadata:
> labels:
> spark-app-id: my-custom-id
> annotations:
> spark-driver-cpu: 1
> spark-driver-mem: 1
> spark-executor-cpu: 1
> spark-executor-mem: 1
> spark-executor-count: 1
> spec:
> schedulerName: spark-scheduler{code}
> Executor Template
>
> {code:java}
> apiVersion: v1
> kind: Pod
> metadata:
> labels:
> spark-app-id: my-custom-id
> spec:
> schedulerName: spark-scheduler{code}
> Kubernetes Pods Launched - Two Executor Pods were launched which was default
> {code:java}
> spark-pi-e608e7718f11cc69-driver 1/1 Running 0 10s
> spark-pi-e608e7718f11cc69-exec-1 1/1 Running 0 5s
> spark-pi-e608e7718f11cc69-exec-2 1/1 Running 0 5s{code}
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-31482) spark.kubernetes.driver.podTemplateFile Configuration not used by the job
[
https://issues.apache.org/jira/browse/SPARK-31482?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17087150#comment-17087150
]
Ed Mitchell commented on SPARK-31482:
-
I'm not sure where the bug is here TBH. I don't see anything in the
documentation or the Spark code that implies that setting annotations on the
Driver Pod is a valid substitute for setting "spark.executor.memory",
"spark.executor.instances", or "spark.executor.cores" on the Spark Submit
command.
> spark.kubernetes.driver.podTemplateFile Configuration not used by the job
> -
>
> Key: SPARK-31482
> URL: https://issues.apache.org/jira/browse/SPARK-31482
> Project: Spark
> Issue Type: Bug
> Components: Kubernetes
>Affects Versions: 3.0.0
>Reporter: Pradeep Misra
>Priority: Blocker
>
> Spark 3.0 - Running Spark Submit as below and point to a MinKube cluster
> {code:java}
> bin/spark-submit \
> --master k8s://https://192.168.99.102:8443 \
> --deploy-mode cluster \
> --name spark-pi \
> --class org.apache.spark.examples.SparkPi \
> --conf spark.kubernetes.driver.podTemplateFile=../driver_1E.template \
> --conf spark.kubernetes.executor.podTemplateFile=../executor.template \
> --conf spark.kubernetes.container.image=spark:spark3 \
> local:///opt/spark/examples/jars/spark-examples_2.12-3.0.0-preview2.jar 1
> {code}
>
> Spark Binaries - spark-3.0.0-preview2-bin-hadoop2.7.tgz
> Driver Template -
> {code:java}
> apiVersion: v1
> kind: Pod
> metadata:
> labels:
> spark-app-id: my-custom-id
> annotations:
> spark-driver-cpu: 1
> spark-driver-mem: 1
> spark-executor-cpu: 1
> spark-executor-mem: 1
> spark-executor-count: 1
> spec:
> schedulerName: spark-scheduler{code}
> Executor Template
>
> {code:java}
> apiVersion: v1
> kind: Pod
> metadata:
> labels:
> spark-app-id: my-custom-id
> spec:
> schedulerName: spark-scheduler{code}
> Kubernetes Pods Launched - Two Executor Pods were launched which was default
> {code:java}
> spark-pi-e608e7718f11cc69-driver 1/1 Running 0 10s
> spark-pi-e608e7718f11cc69-exec-1 1/1 Running 0 5s
> spark-pi-e608e7718f11cc69-exec-2 1/1 Running 0 5s{code}
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
