You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hadoop.apache.org by Jaret Flores <ja...@gmail.com> on 2016/06/27 13:17:06 UTC
ResourceManager Web UI strange behavior
To practice using Hadoop, I have recently setup 3 VMs each of which were
cloned from a VM having ArchLinux+Hadoop installed. After cloning, i
assigned each VM a hostname and a reserved IP on my LAN:
namenode (10.0.0.3), resourcemanager (10.0.0.4), datanode1 (10.0.0.5)
`fs.defaultFS == hdfs://namenode:9000`
`yarn.resourcemanager.hostname == resourcemanager`
And all VMs are listed in the 'slaves' file
After `sbin/start-dfs.sh` from `namenode` and `sbin/start-yarn.sh` from
`resourcemanager`, all processes are up (seen using `jps`) and the logs on
each machine show no errors. I am able to visit namenode:50070 from my
host machine browser to see the NameNode Web UI and see that I have 3
active DataNodes.
**The Problem**: I am unable to view the ResourceManager Web UI at
resourcemanager:8088. Here is the behavior
Host machine:
* Navigating to resourcemanager:8088 in Chrome gives
`net::ERR_CONNECTION_REFUSED`
* `curl resourcemanager:8088` gives `curl: (7) Failed to connect to
resourcemanager port 8088: Connection refused`
From any guest VM:
* `curl resourcemanager:8088` gives an HTTP 302 (redirect) with `Location:
http://resourcemanager:8088/cluster`.
* Following the redirect, `curl -L resourcemanager:8088` *works* and
returns the HTML.
Specs:
Host Machine OS: Windows 10
VMs OS: ArchLinux (latest, just downloaded this weekend)
I first created a VM, installed ArchLinux+Hadoop and made initial
configurations. All VMs are created/cloned/manager via Hyper-V, all have
access to internet and can SSH between each other.
---
Any ideas? I am a bit stumped since the NameNode Web UI works just fine.
I can't tell if this is just a networking issue, or there is some
difference between the webapps. Anyone have an issue like this before?
Thanks for the help!
-Jaret