3

I am new to Kafka and Hadoop technologies. I was trying to install and run my first Single Node, Single Broker Cluster on AWS EC2 VM instance, I am done with:

1) java installation

2) updating ~/.bashrc and ~/.nash_profile files with java related entries

3) Able to run internal zookeeper instance, but

4) As soon as I am trying to start kafka broker it is throwing me following error message:

$ bin/kafka-server-start.sh config/server.properties
Java HotSpot(TM) 64-Bit Server VM warning: INFO: os::commit_memory(0x00000000c0130000, 986513408, 0) failed; error='Cannot allocate memory' (errno=12)

There is insufficient memory for the Java Runtime Environment to continue.
Native memory allocation (malloc) failed to allocate 986513408 bytes for committing reserved memory.
An error report file with more information is saved as:
/usr/local/kafka/hs_err_pid2549.log

I am not sure what wrong am I doing. This AWS EC2 VM instance is a newly created Ubuntu-t2.micro instance with 8 GB general purpose SSD volume.

halfer
  • 19,824
  • 17
  • 99
  • 186
Chauhan B
  • 461
  • 8
  • 27

3 Answers3

8

It was a memory issue. I updated the value for minimum and maximum memory allocation in kafka-server-startup.sh and zookeeper-server-start.sh as export KAFKA_HEAP_OPTS="-Xmx256M -Xms128M" and started zookeeper and kafka servers both worked fine. Thanks

Chauhan B
  • 461
  • 8
  • 27
7

So I believe the error that you're getting is about RAM memory, not SSD memory.

I did a quick google search and it seems like the t2.micro account only has 1gb of RAM? Is that correct? From the error, and from this Stack Overflow Answer, it seems as though you don't have enough RAM. It says malloc requested 986mb or memory, which is basically 1gb and you only have 1gb on your machine.

Community
  • 1
  • 1
Morgan Kenyon
  • 3,072
  • 1
  • 26
  • 38
  • 3
    Yes, it was a memory issue. I updated the value for minimum and maximum memory allocation in **kafka-server-startup.sh and zookeeper-server-start.sh** as **export KAFKA_HEAP_OPTS="-Xmx256M -Xms128M"** and started zookeeper and kafka servers both worked fine. Thanks. – Chauhan B Aug 17 '15 at 05:21
  • To add to this, you can run systemctl status kafka to see if you do indeed have a memory issue – Prithvi Boinpally Jun 25 '21 at 20:25
0

If the same change needs to be done for Confluent Kafka. For centos, modify the /usr/lib/systemd/system/confluent-kafka.service file and pass the "KAFKA_HEAP_OPTS=-Xmx512M -Xms512M" values in that file. This protects environment from upgrade(s). And then systemctl daemon-reload. Check the values of the new kafka by running status on the service systemctl status confluent-kafka.

RootUser RD
  • 21
  • 2