Hey guys, Is there a way to connect Pinot to a sec...
# general
s
Hey guys, Is there a way to connect Pinot to a secure Kafka cluster using Jaas configuration?
m
@User might have some insights here.
a
Hi Satyam, yes. I remember @User was settings up something similar some time ago. If you look up his posts via search you’ll find a number of actual configs
s
@User can you please help find it?
m
@User ^^
m
should be only one change to get it to work with SSL
which is described in the copy
Here's one showing Kafka with SSL & SASL using Jaas - https://dev.startree.ai/docs/pinot/recipes/kafka-ssl-sasl
s
Thanks @User Will this work on the open source release of pinot?
m
That's what I've been using, so yup!
s
Getting this error in the controller logs @User
Copy code
Kafka version: 2.8.1
Kafka commitId: 839b886f9b732b15
Kafka startTimeMs: 1649834197621
[Consumer clientId=consumer-null-648, groupId=null] Subscribed to partition(s): test-topic--2147483648
Metrics scheduler closed
Closing reporter org.apache.kafka.common.metrics.JmxReporter
Metrics reporters closed
App info kafka.consumer for consumer-null-648 unregistered
Transient Exception: Could not get partition count for topic test-topic
org.apache.pinot.spi.stream.TransientConsumerException: org.apache.kafka.common.errors.TimeoutException: Timeout expired while fetching topic metadata
	at org.apache.pinot.plugin.stream.kafka20.KafkaStreamMetadataProvider.fetchPartitionCount(KafkaStreamMetadataProvider.java:50) ~[pinot-kafka-2.0-0.10.0-SNAPSHOT-shaded.jar:0.10.0-SNAPSHOT-078c711d35769be2dc4e4b7e235e06744cf0bba7]
	at org.apache.pinot.spi.stream.StreamMetadataProvider.computePartitionGroupMetadata(StreamMetadataProvider.java:66) ~[pinot-all-0.10.0-SNAPSHOT-jar-with-dependencies.jar:0.10.0-SNAPSHOT-078c711d35769be2dc4e4b7e235e06744cf0bba7]
	at org.apache.pinot.spi.stream.PartitionGroupMetadataFetcher.call(PartitionGroupMetadataFetcher.java:69) ~[pinot-all-0.10.0-SNAPSHOT-jar-with-dependencies.jar:0.10.0-SNAPSHOT-078c711d35769be2dc4e4b7e235e06744cf0bba7]
	at org.apache.pinot.spi.stream.PartitionGroupMetadataFetcher.call(PartitionGroupMetadataFetcher.java:31) ~[pinot-all-0.10.0-SNAPSHOT-jar-with-dependencies.jar:0.10.0-SNAPSHOT-078c711d35769be2dc4e4b7e235e06744cf0bba7]
	at org.apache.pinot.spi.utils.retry.BaseRetryPolicy.attempt(BaseRetryPolicy.java:50) ~[pinot-all-0.10.0-SNAPSHOT-jar-with-dependencies.jar:0.10.0-SNAPSHOT-078c711d35769be2dc4e4b7e235e06744cf0bba7]
	at org.apache.pinot.controller.helix.core.PinotTableIdealStateBuilder.getPartitionGroupMetadataList(PinotTableIdealStateBuilder.java:166) ~[pinot-all-0.10.0-SNAPSHOT-jar-with-dependencies.jar:0.10.0-SNAPSHOT-078c711d35769be2dc4e4b7e235e06744cf0bba7]
m
that error is when you try to create the table?
s
yes
m
hmmm interesting. In the example that I shared with you I had a Confluent Cloud Kafka cluster connected to a local Pinot
so my broker was like this:
Copy code
"stream.kafka.broker.list": "blahblah.us-west-2.aws.confluent.cloud:9092",
and then I put in the consumer key and password, but I didn't do anything else
s
yeah i did the same…
m
I remember seeing this before though
and when I saw it before the problem was that there weren't any messages on partition 0
of the topic
s
In this case, lots of messages are there.
m
ok, so not that
s
Is this a connectivity issue?
have shared schema and ingestion conf with @User
m
ok, hopefully Kartik can see why it's not working. If that doesn't work I could always share my config with you to help narrow down the problem
k
Thanks Mark. Working with Satyam to resolve both the JaaS and Spark ingestion issue.