Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Jul 14 2017 09:41

    pomadchin on master

    Update README.md (compare)

  • Feb 03 2017 20:24
    pomadchin closed #4
  • Feb 03 2017 20:24
    pomadchin commented #4
  • Feb 03 2017 20:23

    pomadchin on master

    Fix quay.io link Merge pull request #5 from jwal… (compare)

  • Feb 03 2017 20:23
    pomadchin closed #5
  • Feb 03 2017 20:23
    pomadchin commented #5
  • Feb 03 2017 20:10
    jwalgran opened #5
  • Dec 10 2016 21:50
    pomadchin commented #4
  • Dec 10 2016 13:42
    Plippe opened #4
Rakesh
@rakeshkr00
After deployment of the docker images with Accumulo 1.9.2, I am encountering below error. Any help would be appreciated.
docker logs accumulo-master
Template config: /etc/hadoop/conf/core-site.xml from /etc/hadoop/conf/core-site.xml.template
/sbin/hdfs-lib.sh: line 11: HADOOP_MASTER_ADDRESS: unbound variable
Found config: /etc/hadoop/conf/core-site.xml
/etc/hadoop/conf/core-site.xml:1: parser error : Document is empty
Grigory
@pomadchin
hey @rakeshkr00 so you don’t have HADOP_MASTER_ADDRESS var set?
Rakesh
@rakeshkr00
@pomadchin where to set that env variable? On EMR I bootstrap using the bootstrap script given in the geomesa-accumulo.
Grigory
@pomadchin
Anywhere; this var has to be passed on the container startup
^ just a docker-compose example but can give you clues about how to start it
Rakesh
@rakeshkr00
In EMR, it asks for the bootstrap script and then it prepare and start the cluster.
Grigory
@pomadchin
just pass it via -e VARNAME=VARVALUE
Rakesh
@rakeshkr00
@pomadchin only after the cluster instances are created, I could know the HADOOP_MASTER_ADDRESS. The bootstrap script is given as part of request for an EMR cluster. So, I don't see how I can provide the -e VARNAME=VARVALUE
Grigory
@pomadchin
hm it is possible
i just don’t remember how
i’ll get back to you soon
MASTER_IP=$(xmllint --xpath "//property[name='yarn.resourcemanager.hostname']/value/text()"  /etc/hadoop/conf/yarn-site.xml)
^ you can do smth like that in boostrap script
Rakesh
@rakeshkr00
Sure. Thanks
Rakesh
@rakeshkr00
thanks @pomadchin I got past that error using your suggestion.
Grigory
@pomadchin
:+1:
Rakesh
@rakeshkr00

When I am connecting to accumulo shell it gives

ERROR: Problem reading instance id out of hdfs at hdfs://<hostname>:8020/gis/instance_id

why it is not picking up from https://github.com/geodocker/geodocker-geomesa/blob/master/geodocker-accumulo-geomesa/docker-compose.yml. Pls help.

the port in the error msg is 8020 and in docker-compose it is 8121
Grigory
@pomadchin
I think you can specify shell port
Are you are trying to log into the shell from your accumulo master container?
Rakesh
@rakeshkr00
@pomadchin to me it seems like for some reason docker-compose.yml is not getting executed/picked up. Previously I had issue with HADOOP_MASTER_ADDRESS (which I got past after using -e ENVVAR=VARVALUE), and now the zookeeper port issue.
Grigory
@pomadchin
Accumulo servers are up?
It’s UI is available?
Rakesh
@rakeshkr00
the full error was
2018-11-20 20:56:47,718 [zookeeper.ZooUtil] ERROR: Problem reading instance id out of hdfs at hdfs://<hostName>:8020/gis/instance_id
org.apache.hadoop.security.AccessControlException: Permission denied: user=root, access=EXECUTE, inode="/gis":hadoop:hadoop:drwx------
Grigory
@pomadchin
ha indeed
looks like smth is wrong with zoo
ah w8
i think we need geomesa people there to help
smth with the --hostname of containers there
looks like some hadoop stack security thing
or probably smth is not up
or tries to be up but in a wrong moment
@elahrvivaz
Emilio
@elahrvivaz
is it just the shell that's failing?
seems to be hdfs user permissions
Austin Heyne
@aheyne
probably need to do sudo -u hadoop stuff
Rakesh
@rakeshkr00
it seem permission as geomesa is not installed. geomesa says command not found.
Emilio
@elahrvivaz
the command is geomesa-accumulo now
Rakesh
@rakeshkr00

@elahrvivaz

geomesa-accumulo
Error: Could not find or load main class .usr.lib.zookeeper.zookeeper.jar

Thanks

Emilio
@elahrvivaz
seems like something is screwed up in that script - that shouldn't be the main class being executed
Rakesh
@rakeshkr00
oh
Rakesh
@rakeshkr00
fyi: after giving hadoop hdfs chmod 777 /gis, the accumulo shell came up fine, though 777 is not an advisable permission
Emilio
@elahrvivaz
you could run the shell as the hadoop user like @aheyne suggested
Rakesh
@rakeshkr00
what would be hadoop user password when challenged by accumulo shell ?
Rakesh
@rakeshkr00
geomesa-accumulo issue seems to be major stumbling block.
Emilio
@elahrvivaz
you'd have to sudo to the hadoop user, then run the accumulo shell with -u root
to debug the geomesa-accumulo script, you can edit it so that the line near the end where it invokes java instead just echos the command out
just add an echo to the front of the line
that will hopefully show what is being garbled