You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@accumulo.apache.org by Jonathan Lasko <jl...@bbn.com> on 2015/11/24 21:09:20 UTC

loading iterator jars from file:// or http://

My project is loading custom iterator jars from HDFS and have run into 
some VFS problems (e.g. redeploying the same jar to HDFS causes 
NoClassDefFoundErrors due to NPEs in VFS). While one solution may be to 
weed out and eliminate any such "redeploys," I am curious if anyone has 
used contexts to load jars from file:// or http:// locations?

Thanks,

Jonathan

Re: loading iterator jars from file:// or http://

Posted by dl...@comcast.net.
Suggest replacing the VFS 2.0 jar in the Accumulo lib directory with a VFS 2.1-SNAPSHOT jar from [1]. There are some bugs that have been fixed, but VFS 2.1 has not been released yet. I did this and have not had many issues loading from HDFS since. Loading jars from file:/// should work, I believe some of the Accumulo tests are written that way. 

[1] https://continuum-ci.apache.org/continuum/workingCopy.action?projectId=129&projectName=Apache+Commons+VFS&userDirectory=core/target 


----- Original Message -----

From: "Jonathan Lasko" <jl...@bbn.com> 
To: "accumulo-user" <us...@accumulo.apache.org> 
Sent: Tuesday, November 24, 2015 3:09:20 PM 
Subject: loading iterator jars from file:// or http:// 

My project is loading custom iterator jars from HDFS and have run into 
some VFS problems (e.g. redeploying the same jar to HDFS causes 
NoClassDefFoundErrors due to NPEs in VFS). While one solution may be to 
weed out and eliminate any such "redeploys," I am curious if anyone has 
used contexts to load jars from file:// or http:// locations? 

Thanks, 

Jonathan