I am trying to setup a 6 node DSE 5.1 spark cluster on AWS EC2 machines.
I have referred the DSE documentation
just to start with, I have opened all TCP ports , when I checked the logs, I found that worker process and executor process and driver process are using below ports
33xxx
33xxx
33xxx
34xxx
34xxx
34xxx
35xxx
35xxx
35xxx
36xxx
37xxx
37xxx
39xxx
40xxx
40xxx
41xxx
41xxx
43xxx
43xxx
43xxx
43xxx
45xxx
46xxx
the range here is from 33xxx to 46xxx, what is suggested range to open the ports ?, or is there any way to bind worker and executor ports ?
By default the port selection is random
See the Spark Docs
Specifically
spark.blockManager.port
spark.driver.port
While you can lock these down to a specific value by setting them in the SparkConf or on the CLI through Spark Submit, you need to make sure that every application has unique values so they do not collide.
In most cases it makes sense to keep the Driver in the same VPN as the Cluster.
Related
Specs:
The serverless Amazon MSK that's in preview.
t2.xlarge EC2 instance with Amazon Linux 2
Installed Kafka from https://dlcdn.apache.org/kafka/3.0.0/kafka_2.13-3.0.0.tgz
openjdk version "11.0.13" 2021-10-19 LTS
OpenJDK Runtime Environment 18.9 (build 11.0.13+8-LTS)
OpenJDK 64-Bit Server VM 18.9 (build 11.0.13+8-LTS, mixed mode,
sharing)
Gradle 7.3.3
https://github.com/aws/aws-msk-iam-auth, successfully built.
I also tried adding IAM authentication information, as recommended by the Amazon MSK Library for AWS Identity and Access Management. It says to add the following in config/client.properties:
# Sets up TLS for encryption and SASL for authN.
security.protocol = SASL_SSL
# Identifies the SASL mechanism to use.
sasl.mechanism = AWS_MSK_IAM
# Binds SASL client implementation.
# sasl.jaas.config = software.amazon.msk.auth.iam.IAMLoginModule required;
# Encapsulates constructing a SigV4 signature based on extracted credentials.
# The SASL client bound by "sasl.jaas.config" invokes this class.
sasl.client.callback.handler.class = software.amazon.msk.auth.iam.IAMClientCallbackHandler
# Binds SASL client implementation. Uses the specified profile name to look for credentials.
sasl.jaas.config = software.amazon.msk.auth.iam.IAMLoginModule required awsProfileName="kafka-client";
And kafka-client is the IAM role attached to the EC2 instance as an instance profile.
Networking: I used VPC Reachability Analyzer to confirm that the security groups are configured correctly and the EC2 instance I'm using as a Producer can reach the serverless MSK cluster.
What I'm trying to do: create a topic.
How I'm trying: bin/kafka-topics.sh --create --partitions 1 --replication-factor 1 --topic quickstart-events --bootstrap-server boot-zclcyva3.c2.kafka-serverless.us-east-2.amazonaws.com:9098
Result:
Error while executing topic command : Timed out waiting for a node assignment. Call: createTopics
[2022-01-17 01:46:59,753] ERROR org.apache.kafka.common.errors.TimeoutException: Timed out waiting for a node assignment. Call: createTopics
(kafka.admin.TopicCommand$)
I'm also trying: with the plaintext port of 9092. (9098 is the IAM-authentication port in MSK, and serverless MSK uses IAM authentication by default.)
All the other posts I found on SO about this node assignment error didn't include MSK. I tried suggestions like uncommenting the listener setting in server.properties, but that didn't change anything.
Installing kcat for troubleshooting didn't work for me, since there's no out-of-the box installation for the yum package manager, which Amazon Linux 2 uses, and since these instructions failed for me at checking for libcurl (by compile)... failed (fail).
The Question: Any other tips on solving this "node assignment" error?
The documentation has been updated recently, I was able to follow it end to end without any issue (The IAM policy is now correct)
https://docs.aws.amazon.com/msk/latest/developerguide/serverless-getting-started.html
The created properties file is not automatically used; your command needs to include --command-config client.properties, where this properties file is documented at the MSK docs on the linked IAM page.
Extract...
ssl.truststore.location=<PATH_TO_TRUST_STORE_FILE>
security.protocol=SASL_SSL
sasl.mechanism=AWS_MSK_IAM
sasl.jaas.config=software.amazon.msk.auth.iam.IAMLoginModule required;
sasl.client.callback.handler.class=software.amazon.msk.auth.iam.IAMClientCallbackHandler
Alternatively, if the plaintext port didn't work, then you have other networking issues
Beyond these steps, I suggest reaching out to MSK support, and telling them to update the "Create a Topic" page to no longer use Zookeeper, keeping in mind that Kafka 3.0 is not (yet) supported
I am trying to setup kafka with 3 broker nodes and 1 zookeeper node in AWS EC2 instances. I have following server.properties for every broker:
kafka-1:
broker.id=0
listeners=PLAINTEXT_1://ec2-**-***-**-17.eu-central-1.compute.amazonaws.com:9092
advertised.listeners=PLAINTEXT_1://ec2-**-***-**-17.eu-central-1.compute.amazonaws.com:9092
listener.security.protocol.map=,PLAINTEXT_1:PLAINTEXT
inter.broker.listener.name=PLAINTEXT_1
zookeeper.connect=ec2-**-***-**-105.eu-central-1.compute.amazonaws.com:2181
kafka-2:
broker.id=1
listeners=PLAINTEXT_2://ec2-**-***-**-43.eu-central-1.compute.amazonaws.com:9093
advertised.listeners=PLAINTEXT_2://ec2-**-***-**-43.eu-central-1.compute.amazonaws.com:9093
listener.security.protocol.map=,PLAINTEXT_2:PLAINTEXT
inter.broker.listener.name=PLAINTEXT_2
zookeeper.connect=ec2-**-***-**-105.eu-central-1.compute.amazonaws.com:2181
kafka-3:
broker.id=2
listeners=PLAINTEXT_3://ec2-**-***-**-27.eu-central-1.compute.amazonaws.com:9094
advertised.listeners=PLAINTEXT_3://ec2-**-***-**-27.eu-central-1.compute.amazonaws.com:9094
listener.security.protocol.map=,PLAINTEXT_3:PLAINTEXT
inter.broker.listener.name=PLAINTEXT_3
zookeeper.connect=ec2-**-***-**-105.eu-central-1.compute.amazonaws.com:2181
zookeeper:
tickTime=2000
dataDir=/var/lib/zookeeper
clientPort=2181
When I ran following command in zookeeper I see that they are connected
I also telnetted from any broker to other ones with broker port they are all connected
However, when I try to create topic with 2 replication factor I get Timed out waiting for a node assignment
I cannot understand what is incorrect with my setup, I see 3 nodes running in zookeeper, but having problems when creating topic. BTW, when I make replication factor 1 I get the same error. How can I make sure that everything is alright with my cluster?
It's good that telnet checks the port is open, but it doesn't verify the Kafka protocol works. You could use kcat utility for that, but the fix includes
listeners are set to either PLAINTEXT://:9092 or PLAINTEXT://0.0.0.0:9092 for every broker, which means using the same port
Removing the number from the listener mapping and advertised listeners property such that each broker is the same
I'd also recommend looking at using Ansible/Terraform/Cloudformation to ensure you consistently modify the cluster rather than edit individual settings manually
Installing wso2am-analytics-2.2.0 on the port offset 0, then I get error messages as
WARN {org.apache.spark.scheduler.TaskSetManager} - Lost task 0.0 in stage 2990.0 (TID 147439, 10.0.11.26): FetchFailed(BlockManagerId(0, someserver.compute.internal, 12001), shuffleId=745, mapId=0, reduceId=0, message=
org.apache.spark.shuffle.FetchFailedException: Failed to connect to ip-10-0-17-131.eu-central-1.compute.internal:12001
Apparently somewhere is configured to connect to port 12001 (while seems the server listens on 12000)
Where could I configure the port 12000?
Thanks
This port is defined in <Product_Home>repository/conf/analytics/spark/spark-defaults.conf. Property name is spark.blockManager.port. However you shouldn't manually configure it.
This particular issue is a connectivity problem in my knowledge. DAS uses 1200x range ports to spark executor communications. So incase of multiple executors or new executor spawning in and event of one executor getting killed incremented port will be opened. Hence at the network level also we should allow traffic through that port range. So opening that port range in your network interface ip-10-0-17-131.eu-central-1.compute.internal will solve your issue.
I am running spark 2.0.1 using below options in
SparkSession.builder().master(master).appName(appName).config(conf).getOrCreate();
opts.put("spark.serializer","org.apache.spark.serializer.KryoSerializer");
opts.put("spark.executor.extraJavaOptions","-verbose:gc -XX:+PrintGCDetails -XX:+PrintGCTimeStamps -XX:+UseG1GC -Djava.security.egd=file:///dev/urandom");
opts.put("spark.driver.maxResultSize","0");
opts.put("spark.sql.shuffle.partitions","200");
opts.put("spark.sql.warehouse.dir","/opt/astra/spark-warehouse");
opts.put("spark.scheduler.mode","FAIR");
opts.put("spark.executor.memory","5g");
opts.put("spark.executor.cores","2");
opts.put("spark.kryoserializer.buffer.max","1g");
opts.put("spark.parquet.block.size","134217728");
Spark master is running in AWS on ec2 instance. In spark master ui I can see memory, cores all. But when running job in Job UI executors as below
Also while looking at thread dump I see lots of connection related threads awaiting.
Can someone please point to me what's happening and where to look. As commented here is spark master's snapshot showing allocated resources.
On logs the system seems waiting for resources providing below link
16/11/08 12:46:37 WARN TaskSchedulerImpl: Initial job has not accepted any resources; check your cluster UI to ensure that workers are registered and have sufficient resources
I am trying to execute pyspark from my mac to do compute on a EC2 spark cluster.
If I login to the cluster, it works as expected:
$ ec2/spark-ec2 -i ~/.ec2/spark.pem -k spark login test-cluster2
$ spark/bin/pyspark
Then do a simple task
>>> data=sc.parallelize(range(1000),10)`
>>> data.count()
Works as expected:
14/06/26 16:38:52 INFO spark.SparkContext: Starting job: count at <stdin>:1
14/06/26 16:38:52 INFO scheduler.DAGScheduler: Got job 0 (count at <stdin>:1) with 10 output partitions (allowLocal=false)
14/06/26 16:38:52 INFO scheduler.DAGScheduler: Final stage: Stage 0 (count at <stdin>:1)
...
14/06/26 16:38:53 INFO spark.SparkContext: Job finished: count at <stdin>:1, took 1.195232619 s
1000
But now if I try the same thing from local machine,
$ MASTER=spark://ec2-54-234-204-13.compute-1.amazonaws.com:7077 bin/pyspark
it can't seem to connect to the cluster
14/06/26 09:45:43 INFO AppClient$ClientActor: Connecting to master spark://ec2-54-234-204-13.compute-1.amazonaws.com:7077...
14/06/26 09:45:47 WARN TaskSchedulerImpl: Initial job has not accepted any resources; check your cluster UI to ensure that workers are registered and have sufficient memory
...
File "/Users/anthony1/git/incubator-spark/python/lib/py4j-0.8.1-src.zip/py4j/protocol.py", line 300, in get_return_value
py4j.protocol.Py4JJavaError: An error occurred while calling o20.collect.
: org.apache.spark.SparkException: Job aborted: Spark cluster looks down
14/06/26 09:53:17 WARN TaskSchedulerImpl: Initial job has not accepted any resources; check your cluster UI to ensure that workers are registered and have sufficient memory
I thought the problem was in the ec2 security but it does not help even after adding inbound rules to both master and slave security groups to accept all ports.
Any help will be greatly appreciated!
Others are asking same question on mailing list
http://apache-spark-user-list.1001560.n3.nabble.com/Deploying-a-python-code-on-a-spark-EC2-cluster-td4758.html#a8465
The spark-ec2 script configure the Spark Cluster in EC2 as standalone, which mean it can not work with remote submits. I've been struggled with this same error you described for days before figure out it's not supported. The message error is unfortunately incorrect.
So you have to copy your stuff and log into the master to execute your spark task.
In my experience Initial job has not accepted any resources; check your cluster UI to ensure that workers are registered and have sufficient memory usually means you have accidentally set the cores too high, or set the executer memory too high - i.e. higher than what your nodes actually have.
Other, less likely causes, could be you got the URI wrong and your not really connecting to the master. And once I saw that problem when the /run partition was 100%.
Even less likely, your cluster may actually be down, and you need to restart your spark workers.