Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Threaded View
Accumulo >> mail # user >> GLIBC error


Copy link to this message
-
Re: GLIBC error
(Even though jvines beat me to it...)

Did you build the Accumulo distribution on the node which you're having
these troubles on?

The native maps (the libNativeMap-Linux-amd64-64.so file) is C++ code
which is called through JNDI. I'm rusty on my dynamic linking, but check
the version of gcc and glibc installed on the host in question. I'm
guessing there person who built the dist tarball you downloaded doesn't
match the versions you have on that host. Post your linux distro/release
and the output of `gcc -v` if you need more guidance.

An alternative, although not a very desirable one, is to move the native
maps out of lib/native (just move it to lib/native-orig and the tserver
will just use Java code for the inMemoryMaps. Your performance will
suffer, but the tserver should start right up.

On 07/24/2012 04:21 PM, Cardon, Tejay E wrote:
>
> I�ve been trying to run Accumulo on a new cluster, but I�m having
> trouble. I�ve done this many times before, but this time I�m getting
> an error on the tservers regarding glibc. I think I�ve tracked it down
> to a version issue (the version of Glibc I�ve got installed isn�t new
> enough, so it doesn�t contain the right version of GLIBCXX). I ran
> into this once before, and the solution was to build Accumulo from
> scratch with Maven. However, I�ve tried that this time with no
> success. If I download the *dist.tar.gz and do the build from there,
> it fails. If I download the *-src.tar.gz and build from it, I can�t
> seem to find the final packaged .tar.gz. If I use the src.tar.gz,
> build it, and then just use the directory I built from, I get an error
> about setting up log4j. Any suggestions?
>
> Thanks
>
> Tejay
>
> Here�s the error I get if I use the �dist.tar.gz This error is found
> in the .debug.log and .log files from the tserver.
>
> ERROR: Failed to load native map library
> /opt/accumulo-1.4.0/lib/native/map/libNativeMap-Linux-amd64-64.so
>
> java.lang.UnsatisfiedLinkError:
> /opt/accumulo-1.4.0/lib/native/map/libNativeMap-Linux-amd64-64.so:
> /usr/lib64/libstdc++.so.6: version `GLIBCXX_3.4.9' not found (required
> by /opt/accumulo-1.4.0/lib/native/map/libNativeMap-Linux-amd64-64.so)
>
> at java.lang.ClassLoader$NativeLibrary.load(Native Method)
>
> at java.lang.ClassLoader.loadLibrary0(ClassLoader.java:1807)
>
> at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1703)
>
> at java.lang.Runtime.load0(Runtime.java:770)
>
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB