Hi <@UDRLN8MAP> I'm working with <@U01G51G5VCG> an...
# general
j
Hi @User I'm working with @User and trying to use our new Pinot Kinesis support. Want to know do we have any images built on that? Because with the branch, we can not use it directly. Thanks
n
we haven’t built any images. You could build one from the branch: https://docs.pinot.apache.org/operators/tutorials/build-docker-images
j
Thanks Neha, I tried it according to the doc and got the error
[ERROR] Failed to execute goal com.mycila:license-maven-plugin:2.8:check (default) on project pinot-kinesis: Some files do not have the expected license header -> [Help 1]
[ERROR]
 
[ERROR] To see the full stack trace of the errors, re-run Maven with the -e switch.
[ERROR] Re-run Maven using the -X switch to enable full debug logging.
[ERROR]
 
[ERROR] For more information about the errors and possible solutions, please read the following articles:
[ERROR] [Help 1] <http://cwiki.apache.org/confluence/display/MAVEN/MojoExecutionException>
[ERROR]
 
[ERROR] After correcting the problems, you can resume the build with the command
[ERROR]  mvn <args> -rf :pinot-kinesis
The command '/bin/sh -c git clone ${PINOT_GIT_URL} ${PINOT_BUILD_DIR} &&   cd ${PINOT_BUILD_DIR} &&   git checkout ${PINOT_BRANCH} &&   mvn install package -DskipTests -Pbin-dist -Pbuild-shaded-jar -Dkafka.version=${KAFKA_VERSION} &&   mkdir -p ${PINOT_HOME}/configs &&   mkdir -p ${PINOT_HOME}/data &&   cp -r pinot-distribution/target/apache-pinot-*-bin/apache-pinot-*-bin/* ${PINOT_HOME}/. &&   chmod +x ${PINOT_HOME}/bin/*.sh' returned a non-zero code: 1
do you know how to deal with it?
n
looks like some files are missing license headers. I’ll add them and update the branch
thankyou 1
you can try again
j
Hi Neha, I deployed the new image and seems like the UI of Pinot doesn't running by using the new image. Do you have any ideas with that?
n
that is strange. that should not happen.. the kinesis branch is only a few commits behind master, and the UI code was merged long time ago
what exactly do you mean by the UI is not running? what do you see?
j
The UI is a blank page, when I switch back to the official image, the UI come back gain.
n
did you try clearing browser cache ?
j
Just worked! Thanks Neha, but one more question, why brokers all dead with this image. I redeployed few times but brokers keep dead
n
Why do you say they are dead? You’re seeing errors when you try to query? Sometimes the cluster can take a few minutes to be all active
j
Screen Shot 2021-01-14 at 3.32.37 PM.png
n
hmm, can you check broker logs? are servers alive?
j
In the broker logs, there is an error like this
Copy code
2021/01/14 17:10:27.612 WARN [ClientCnxn] [Start a Pinot [BROKER]-SendThread(wb-metrics-zookeeper.wb-metrics.svc.cluster.local:2181)] Session 0x0 for server null, unexpected error, closing socket connection and attempting reconnect
java.net.ConnectException: Connection refused
	at sun.nio.ch.SocketChannelImpl.checkConnect(Native Method) ~[?:1.8.0_275]
	at sun.nio.ch.SocketChannelImpl.finishConnect(SocketChannelImpl.java:716) ~[?:1.8.0_275]
	at org.apache.zookeeper.ClientCnxnSocketNIO.doTransport(ClientCnxnSocketNIO.java:361) ~[pinot-all-0.7.0-SNAPSHOT-jar-with-dependencies.jar:0.7.0-SNAPSHOT-efbeea2bf8159eb7299292929a57c47345feb341]
	at org.apache.zookeeper.ClientCnxn$SendThread.run(ClientCnxn.java:1144) [pinot-all-0.7.0-SNAPSHOT-jar-with-dependencies.jar:0.7.0-SNAPSHOT-efbeea2bf8159eb7299292929a57c47345feb341]
n
so brokers are not able to connect to Zookeeper? Are other components able to start up fine like servers and controller?
@Daniel Lavoie @Xiang Fu could one of you help to investigate this one?
j
Yes, others are able to start up fine
x
hmm can you try to delete the broker pods and see it can fix the issue
this is a warning log should reconnect to zk
j
Yes, i tried deleting them, when they start again, still shown dead in the UI
x
Hmm, any thing found in broker logs
j
So wired the logs looks good.
Copy code
2021/01/15 18:28:51.181 INFO [HelixBrokerStarter] [Start a Pinot [BROKER]] Starting Pinot broker
2021/01/15 18:28:51.189 INFO [HelixBrokerStarter] [Start a Pinot [BROKER]] Connecting spectator Helix manager
2021/01/15 18:28:52.241 INFO [HelixBrokerStarter] [Start a Pinot [BROKER]] Setting up broker request handler
2021/01/15 18:28:54.144 INFO [HelixBrokerStarter] [Start a Pinot [BROKER]] Starting broker admin application on port: 8099
Jan 15, 2021 6:28:58 PM org.glassfish.grizzly.http.server.NetworkListener start
INFO: Started listener bound to [0.0.0.0:8099]
Jan 15, 2021 6:28:58 PM org.glassfish.grizzly.http.server.HttpServer start
INFO: [HttpServer] Started.
x
hmm
can you check controller zookeeper browser and see what’s under LIVEINSTANCES?
image.png
are those brokers joined as expected
j
No, they are not showing there
x
hmm, are those broker containers are running ? can you enter into the container and see the pinotBroker.log file
that should give extra logs than kubectl logs
j
Hi @Neha Pawar I just debugged with Xiang and seems there are some problems in the kinesis branch. Because when switch back to the original image, it works well.
n
did you figure out what’s the exact problem? was there nothing in the logs?
@Xiang Fu could some problem have come up because of the jackson version upgrade? that is the only thing in that branch which would affect brokers
x
could be
but we don’t see anything from the log
i feel you can try to start controller/broker/server separately
and see if that can be reproduced
n
i’m able to start all components separately, locally just fine. I’ve been testing it that way using StartZK, StartController StartBroker and StartServer locally
x
hmm
which branch are you using, I can try to build one docker image using that
n
sharded_consumer_type_support_with_kinesis
x
@Jinwei Zhu can you try to deploy this image:
apachepinot/pinot:sharded_consumer_type_support_with_kinesis
I built it from the branch neha provided
n
it worked for you @Xiang Fu ?
j
Hmm, seems the same as before.😔
Any updates?🙂
n
Xiang suggested you use the master branch for brokers and the kinesis branch for everything else right? did you try that?
can you operate in that mode until you are testing? we will investigate this before publishing kinesis stuff to master
j
Got it. Thanks
Any ideas on how to do that? should I only redeploy all the brokers using some broker images?
d
You can use
kubectl edit pod
to change the container image tag to latest for broker.
x
kubectl edit statefulset 🙂
d
you got me 🙂
n
were you able to get it running @Jinwei Zhu ?
j
Yeah, with the master branch for brokers, I can run it
n
were you able to setup the kinesis ingestion?