I am trying to run a PySpark job on a Mesosphere cluster but I cannot seem to get it to run. I understand that Mesos does not support cluster deploy mode for PySpark applications and that it needs to be run in client mode. I believe this is where the problem lies.
When I try submitting a PySpark job I am getting the output below.
... socket.hpp:107] Shutdown failed on fd=48: Transport endpoint is not connected [107]
I believe that a spark job running in client mode needs to connect to the nodes directly and this is being blocked?
What configuration would I need to change to be able to run a PySpark job in client mode?
When running PySpark in client mode (meaning the driver is running where you invoke Python) the driver becomes the Mesos Framework. When this happens, the host the framework is running on needs to be able to connect to all nodes in the cluster, and they need to be able to connect back, meaning no NAT.
If this is indeed the cause of your problems, there are two environment variables that might be useful. If you can get a VPN in place, you can set LIBPROCESS_IP and SPARK_LOCAL_IP both to the IP of the host machine that cluster nodes can use to connect back to the driver.
Related
I'm trying to use the following Helm Chart for Spark on Kubernetes
https://github.com/bitnami/charts/tree/main/bitnami/spark
The documentation is of course spotty but I've muddled along. So I have it installed with custom values that assign things like resource limits etc. I'm accessing the master through a NodePort and the WebUI through a port forward. I am NOT using spark-submit, I'm writing Python code to drive the Spark Cluster as follows:
import pyspark
sc = pyspark.SparkContext(appName="Testy", master="spark://<IP>:<PORT>")
This Python code is running locally on my Windows laptop, the Kubernetes cluster is on a separate set of servers. It connects and I can see the app appear in the WebUI but the second it tries to do something I get the following:
WARN TaskSchedulerImpl: Initial job has not accepted any resources; check your cluster UI to ensure that workers are registered and have sufficient resources
The master seems to be in a cycle of removing and launching executors and the 3 workers each just fail to run a launch command. Interestingly the command has the hostname of my laptop in here:
"--driver-url" "spark://CoarseGrainedScheduler#<laptop hostname>:60557"
Got to imagine that's not right. So in this setup where should I be actually running the python code? On the kubernetes cluster? Can I run it locally on my laptop? These details are of course missing from the docs. I'm new to Spark so just looking for the absolute basics. My preferred workflow would be to develop code locally on my laptop then run it on the Kubernetes cluster I have access to.
I know there is information worth 10 google pages on this but, all of them tell me to just put --master yarn in the spark-submit command. But, in cluster mode, how can my local laptop even know what that means? Let us say I have my laptop and a running dataproc cluster. How can I use spark-submit from my laptop to submit a job to this cluster?
Most of the documentation on running a Spark application in cluster mode assumes that you are already on the same cluster where YARN/Hadoop are configured (e.g. you are ssh'ed in), in which case most of the time Spark will pick up the appropriate local configs and "just work".
This is same for Dataproc: if you ssh onto the Dataproc master node, you can just run spark-submit --master yarn. More detailed instructions can be found in the documentation.
If you are trying to run applications locally on your laptop, this is more difficult. You will need to set up an ssh tunnel to the cluster, and then locally create configuration files that tell Spark how to reach the master via the tunnel.
Alternatively, you can use the Dataproc jobs API to submit jobs to the cluster without having to directly connect. The one caveat is that you will have to use properties to tell Spark to run in cluster mode instead of client mode (--properties spark.submit.deployMode=cluster). Note that when submitting jobs via the Dataproc API, the difference between client and cluster mode is much less pressing because in either case the Spark driver will actually run on the cluster (on the master or a worker respectively), not on your local laptop.
In client deploy mode a Spark driver needs to be able to receive incoming TCP connections from Spark executors. However, if the Spark driver is behind a NAT, it cannot receive incoming connections. Will running the Spark driver in YARN cluster deploy mode overcome this limitation of being behind a NAT, because the Spark driver is then apparently executed on the Spark master?
Will running the Spark driver in YARN cluster deploy mode overcome this limitation of being behind a NAT, because the Spark driver is then apparently executed on the Spark master?
Yes, it will. Another possible approach is to configure:
spark.driver.port
spark.driver.bindAddress
and create SSH tunnel to one of the nodes.
I read an answer from What conditions should cluster deploy mode be used instead of client?,
(In client mode) You could run spark-submit on your laptop, and the Driver Program would run on your laptop.
Also, the Spark Doc says,
In client mode, the driver is launched in the same process as the client that submits the application.
Does it mean that I can submit spark tasks from any machine, as long as it can be reachable from master and has Spark environment?
Or in other words, can driver process run outside of the Spark cluster?
Yes, the driver can run on your laptop. Keep in mind though:
The Spark driver will need the Hadoop configuration to be able to talk to YARN and HDFS. You could copy it from the cluster and point to it via HADOOP_CONF_DIR.
The Spark driver will listen on a lot of ports and expect the executors to be able to connect to it. It will advertise the hostname of your laptop. Make sure it can be resolved and all ports accessed from the cluster environment.
Yes, I'm running spark-submit jobs over the LAN using option --deploy-mode cluster. Currently running into this issue however: the server response (json object) isn't very descriptive.
When using spark at client mode (e.g. yarn-client), does the local machine that runs the driver communicates directly with the cluster worker nodes that run the remote executors?
If yes, does it mean the machine (that runs the driver) need to have network access to the worker nodes? So the master node requests resources from the cluster, and returns the IP addresses/ports of the worker nodes to the driver, so the driver can initiating the communication with the worker nodes?
If not, how does the client mode actually work?
If yes, does it mean that the client mode won't work if the cluster is configured in a way that the work nodes are not visible outside the cluster, and one will have to use cluster mode?
Thanks!
The Driver connects to the Spark Master, requests a context, and then the Spark Master passes the Spark Workers the details of the Driver to communicate and get instructions on what to do.
The means that the driver node must be available on the network to the workers, and it's IP must be one that's visible to them (i.e. if the driver is behind NAT, while the workers are in a different network, it won't work and you'll see errors on the workers that they fail to connect to the driver)
When you run Spark in client mode, the driver process runs locally.
In cluster mode, it runs remotely on an ApplicationMaster.
In other words you will need all the nodes to see each other. Spark driver definitely needs to communicate with all the worker nodes. If this is a problem try to use the yarn-cluster mode, then the driver will run inside your cluster on one of the nodes.