apache / accumulo-docker Goto Github PK

Hi,

Ref. the README; will you publish an official image soon?

Docker may be changing registration requirements or process relating to organizations with publishing artifacts in DockerHub. Accumulo docker is likely covered under the Apache umbrella. Check that accumulo-docker is properly registered as an Apache project, and if not, complete the necessary registration process for DockerHub.

In the README.md file, I see the below text

Before you can run Accumulo services in Docker, you will need to install Accumulo, configure accumulo.properties, and initialize your instance with --upload-accumulo-props. This will upload configuration to Zookeeper and limit how much configuration needs to be set on the command line.

```bash
$ accumulo init --upload-accumulo-props
...

So, where is this supposed to be invoked? Are we to run this as docker run command or am I missing something here?

Running a new image,
$> docker run accumulo init . Hangs, warns to be sure Zookeeper is running.
$> docker run accumulo sookeeper . Problem reading instance id out of hdfs at hdfs://localhost:8020/accumulo/instance_id

Do we have a catch 22?

Any reason why no documentation exists for ports that are used and exposed when running this image?

I was following the README to get accumulo running with docker but I was unable to do so without some changes to the Dockerfile. When I ran the Docker build, it would fail to complete the move of hadoop into /opt/hadoop. I was able to get around the issue by switching from moving those dirs to copying them, but that isn't a great solution in the long term. I got into the container at the point just before the move and the file permissions for some the directories under /tmp/hadoop/ were all question marks (???????) and the user and group were just listed as an id.

I think we probably want to add a umask and --no-same-permissions to the tar extract commands to correct that. I will try that out tonight and if that works I will put in a PR.

I built the image and ran docker run accumulo and I got:

HADOOP_HOME=/path/to/hadoop is not set to a valid directory in accumulo-env.sh

I added HADOOP_HOME to Dockerfile and then got:

2018-09-10 21:25:49,371 [start.Main] ERROR: Unable to find Hadoop Configuration class on classpath, check configuration.
java.lang.ClassNotFoundException: org.apache.hadoop.conf.Configuration
	at java.net.URLClassLoader.findClass(URLClassLoader.java:381)
	at java.lang.ClassLoader.loadClass(ClassLoader.java:424)
	at org.apache.accumulo.start.classloader.AccumuloClassLoader$1.loadClass(AccumuloClassLoader.java:198)
	at java.lang.ClassLoader.loadClass(ClassLoader.java:357)
	at org.apache.accumulo.start.Main.main(Main.java:52)

I added ENV HADOOP_CONF_DIR $HADOOP_HOME/etc/hadoop and I still get Unable to find Hadoop Configuration class on classpath, check configuration..