Reputation: 41
I have a 2 node Hadoop cluster running on AWS EC2. I am trying to mount the HDFS on a different EC2 instance. The client is running Ubuntu 12.04.4 LTS and I have installed Hadoop-Fuse.
# apt-cache policy hadoop-0.20-fuse
hadoop-0.20-fuse:
Installed: 0.20.2+923.479-1~maverick-cdh3
Candidate: 0.20.2+923.479-1~maverick-cdh3
Version table:
*** 0.20.2+923.479-1~maverick-cdh3 0
500 http://archive.cloudera.com/debian/ maverick-cdh3/contrib amd64 Packages
100 /var/lib/dpkg/status
After I try to mount it, I get the following:
# hadoop-fuse-dfs dfs://10.0.0.160:9000 /mnt/tmp
INFO fuse_options.c:165 Adding FUSE arg /mnt/tmp
When I run "df" command, I don't see it there and get the input/output error
# df -h
df: `/mnt/tmp': Input/output error
Filesystem Size Used Avail Use% Mounted on
/dev/xvda1 30G 3.5G 25G 13% /
Also the path shows the following:
# ls -alh /mnt
ls: cannot access /mnt/tmp: Input/output error
total 8.0K
drwxr-xr-x 3 root root 4.0K Aug 11 19:42 .
drwxr-xr-x 25 root root 4.0K Aug 11 17:35 ..
d????????? ? ? ? ? ? tmp
Any way I can mount it?
Upvotes: 4
Views: 1724
Reputation: 5957
Unfortunately, hadoop-fuse-dfs
doesn't have good error messages or documentation.
In order for hadoop-fuse-dfs
to work properly you need the Namenode's RPC port which is dfs.namenode.servicerpc-address
in hdfs-site.xml.
# hadoop-fuse-dfs dfs://NAMENODE:RPCPORT /mnt/tmp
Upvotes: 2