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 Plain View
Accumulo >> mail # user >> Accumulo with Hadoop 1.2.0 -- no tablet servers & dfs settings


+
Carlos Mundi 2013-08-04, 02:01
+
Josh Elser 2013-08-04, 02:21
+
Carlos Mundi 2013-08-04, 02:27
Copy link to this message
-
Re: Accumulo with Hadoop 1.2.0 -- no tablet servers & dfs settings
Yes, I believe the value used to be "true".

On 08/03/2013 10:27 PM, Carlos Mundi wrote:
> Hi Josh,
>
> Thanks for your reply!  I set dfs.durable.sync to 1 in my
> hdfs-site.xml like this:
>
> <property>
>   <name>dfs.durable.sync</name>
>   <value>1</value>
> </property>
>
>
> And then I stopped and restarted the whole chain: hadoop, zookeeper,
> accumulo.
>
> Same results!  So I'm not sure what's going on.  Maybe hadoop needs
> 'true' instead of '1'.
>
> I'm tinkering with it and will report any new info.
>
> Cheers!
> Carlos
>
>
>
>
> On Sat, Aug 3, 2013 at 7:21 PM, Josh Elser <[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>> wrote:
>
>     Hi Carlos,
>
>     I pulled down the Apache Hadoop 1.2.0 source release and took a
>     look at src/hdfs/hdfs-default.xml. Reading the description on the
>     dfs.support.append property, dfs.support.append isn't supported
>     and dfs.durable.sync is enabled by default. Looks like we need to
>     change that check in the TabletServer code.
>
>     In the mean time, you can likely set that property in your
>     hdfs-site.xml to be true to fake out the tablet server and make it
>     think that it's set.
>
>     Thanks for letting us know!
>
>     - Josh
>
>
>     On 08/03/2013 10:01 PM, Carlos Mundi wrote:
>
>         Hi.
>
>         This is my first time setting up accumulo.  So I'm starting
>         with a single node "cluster" on Ubuntu 13.04.  My setup is...
>
>         hadoop      1.2.0 (prebuilt tarball from Apache) -- The
>         version will turn out to be important!
>         zookeeper   3.4.5 (prebuilt tarball from Apache)
>         accumulo    1.5.0 (custom build, to target hadoop 1.2)
>
>         As I understand, because I am running hadoop 1.2, I need to do
>         a custom build.  So I did, following the README for Accumulo.
>
>         So I fire up hadoop, then zookeeper and then accumulo. Looks
>         good so far:
>
>         $ jps
>         21231 SecondaryNameNode
>         20733 NameNode
>         24274 Jps
>         22556 Main
>         21994 Main
>         20978 DataNode
>         22306 Main
>         21764 QuorumPeerMain
>         21570 TaskTracker
>         22409 Main
>         21322 JobTracker
>
>         Then I try to fire up an accumulo shell (still following the
>         README) and I am told that there are no tablet servers!  Here
>         is what I get in /usr/local/accumulo/logs/tracer_mybox.log:
>
>         2013-08-03 18:32:44,918 [impl.ServerClient] WARN : There are
>         no tablet servers: check that zookeeper and accumulo are running.
>
>         I check jps, and everything is still running.  And here is
>         what I have in /usr/local/accumulo/logs/tserver_mybox.log:
>
>         2013-08-03 18:32:41,292 [tabletserver.TabletServer] FATAL:
>         Must set dfs.durable.sync OR dfs.support.append to true.
>          Which one needs to be set depends on your version of HDFS.
>          See ACCUMULO-623.
>           HADOOP RELEASE          VERSION           SYNC NAME DEFAULT
>           Apache Hadoop           0.20.205  dfs.support.append false
>           Apache Hadoop            0.23.x dfs.support.append true
>           Apache Hadoop             1.0.x dfs.support.append false
>           Apache Hadoop             1.1.x dfs.durable.sync true
>           Apache Hadoop          2.0.0-2.0.2  dfs.support.append true
>           Cloudera CDH             3u0-3u3             ????          
>           true
>           Cloudera CDH               3u4  dfs.support.append    true
>           Hortonworks HDP           `1.0  dfs.support.append    false
>           Hortonworks HDP           `1.1  dfs.support.append    false
>
>         And, yes, I remember seeing that in the README.  But now you
>         see my problem.  Hadoop 1.2 is not on the list!
>
>         At least, I assume that's my only problem.  :)
>
>         Can someone please tell me which of the two dfs options I need?
>
>         Thanks very much!
>
>         Carlos
>
>
>
+
Carlos Mundi 2013-08-04, 02:49
+
Josh Elser 2013-08-04, 02:55
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