Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Kafka >> mail # dev >> [jira] [Created] (KAFKA-706) broker appears to be encoding ProduceResponse, but never sending it

Copy link to this message
[jira] [Created] (KAFKA-706) broker appears to be encoding ProduceResponse, but never sending it
ben fleis created KAFKA-706:

             Summary: broker appears to be encoding ProduceResponse, but never sending it
                 Key: KAFKA-706
                 URL: https://issues.apache.org/jira/browse/KAFKA-706
             Project: Kafka
          Issue Type: Bug
          Components: core
    Affects Versions: 0.8
         Environment: reproduced on both Mac OS and RH linux, via private node.js client
            Reporter: ben fleis
By all appearances, I seem to be able to convince a broker to periodically encode, but never transmit, a ProduceResponse.  Unfortunately my client is proprietary, but I will share it with Neha via LI channels.  But I will describe what's going on in the hopes that there's another trivial way to reproduce it.  (I did search through JIRA, and haven't found anything that looks like this.)

I am running a single instance zookeeper and single broker.  I have a client that generates configurable amounts of data, tracking what is produced (both sent and ACK'd), and what is consumed.  I was noticing that when using high transfer rates via high frequency single messages, my unack'd queue appeared to be getting continuously larger.  So, I outfitted my client to log more information about correlation ids at various stages, and modified the kafka ProducerRequest/ProducerResponse to log (de)serialization of the same.  I then used tcpdump to intercept all communications between my client and the broker.  Finally, I configured my client to generate 1 message per ~10ms, each payload being approximately 33 bytes; requestAckTimeout was set to 2000ms, and requestAcksRequired was set to 1.  I used 10ms as I found that 5ms or less caused my unacked queue to build up due to system speed -- it simply couldn't keep up.  10ms keeps the load high, but just manageable.  YMMV with that param.  All of this is done on a single host, over loopback.  I ran it on both my airbook, and a well setup RH linux box, and found the same problem.

At startup, my system logged "expired" requests - meaning reqs that were sent, but for which no ACK, positive or negative, was seen from the broker, within 1.25x the requestAckTimeout (ie, 2500ms).  I would let it settle until the unacked queue was stable at or around 0.

What I found is this: ACKs are normally generated within milliseconds.  This was demonstrated by my logging added to the scala ProducerRe* classes, and they are normally seen quickly by my client.  But when the actual error occurs, namely that a request is ignored, the ProducerResponse class *does* encode the correct correlationId; however, a response containing that ID is never sent over the network, as evidenced by my tcpdump traces.  In my experience this would take anywhere from 3-15 seconds to occur after the system was warm, meaning that it's 1 out of several hundred on average that shows the condition.

While I can't attach my client code, I could attach logs; but since my intention is to share the code with LI people, I will wait to see if that's useful here.
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira