AnsweredAssumed Answered

Issues with Mapr Loopback NFS

Question asked by mandoskippy on May 5, 2016
Latest reply on Sep 13, 2016 by leonclayton

I am trying to use the mapr-loopback nfs on M3 with MapR 5.1.0.   Basically, for a small cluster (under 10 nodes) I'd like to be able to install the loopback nfs on each node.  Jim Scott  and his Zeta idea is what got me here, basically to have have a 10 node or less proof of concent cluster using all free tools.  (So having the ability to mount NFS loopback on each node is sorta like running an NFS server on each node, but M3/Community doesn't allow the NFS to run on each node, and with Loopback, I am limited to 10 nodes).

 

Ok, so I have things setup Cluster is up and licensed.  I installed the Loopback client on the nodes, however only one works.... The mount command doesn't provide an an error, there is just no "mycluster" in /mapr on 2 out of 3 nodes.

 

 

The logs on the "not" working nodes have this entry:

 

2016-05-05 19:44:54,1818 INFO loopbacknfs[31244] fs/nfsd/nfsha.cc:478 hostname: ip-10-22-87-236.ec2.internal, hostid: 0x320361225b7d0bc8

2016-05-05 19:44:54,1831 ERROR loopbacknfs[31244] fs/nfsd/nfsha.cc:872 0.0.0.0[0] Error registering with CLDB: Connection reset by peer, err=104, status=0 cldb=10.22.87.235:7222

 

and the logs on the cldb have this:

 

 

2016-05-05 19:47:12,5270 Unable to obtain binding for request from 10.22.87.236:9998. Closing connection

 

 

 

Any thoughts would be appreciated!

 

John

Outcomes