0
votes

We have a server that is handling the production and consumption of messages. We have 4 laptops, all macs with confluent all running the same command line...

./kafka-avro-console-consumer --from-beginning --bootstrap-server 0.0.0.0:9092,0.0.0.0:9092 --topic topicName --property schema.registry.url=http://0.0.0.0:8081

3 of the 4 laptops have no problem consuming the messages, however the fourth will not. I've confirmed that my laptop has the same access to the ip addresses, I'm able to view the topics in the schema registry by going to the address /subjects and I receive no errors in the console nor the log.

I am at a total loss as to what else to check. Has anyone else had that situation and possibly found a solution?

1
How many partitions does the topic you consume have?Arek
It's on our test box so PartitionCount:2 ReplicationFactor:1Peej
I also tried this on our performance server, that one has 48 partitions and I have the same issue. To add more detail, if I run our webproxy in code (c#) which produces a message, it times out after a minute or so. So its both producing and consuming that I'm having issue with.Peej

1 Answers

0
votes

There is a small probability that two (or more) unique console consumer applications get the exact same, "random" consumer group ID as another, and therefore you would only see messages on the instances that were assigned a topic partition that has data.

One solution would be to pass a unique group as a consumer property to every machine, otherwise try restart the command a few times on the machine not getting data to randomize a new group id.

I assume you aren't actually using 0.0.0.0 in the command