Hi, Carl,
Thanks for your efforts on trying reproduce this issue, actually, I've
figured this out inspired by Leo, also my case is little bit more tricky,
it's also due to lack of a compression jar file in hadoop_classpath: our
system use customized compression format in our log, then a customized
Hi Peter,
I was unable to reproduce your problem using a fresh install
of hadoop-0.20.1+169.68 from CDH2 and hive-0.5.0+20 from CDH3. I also
verified that the hadoop-0.20.1+169.68-core.jar contains
org.apache.hadoop.mapred.TextInputFormat. I agree that this is most likely a
classpath issue. My
Hi Ed,
I'm happy to answer questions on this list related to Hive and Cloudera
packaging issues. I'd also like to point out that it's impossible to
differentiate between a cloudera packaging issue and an Apache Hive issue
before you know the cause of the problem.
Thanks.
Carl
On Wed, Sep 15,
This is a me too post: we just ran into an identical problem setting
up a new cluster using CDH3b2. This is all rather mystifying, because
all the correct libraries are there; in fact the hive command line
looks something like
/usr/java/jdk1.6.0_12/bin/java -Xmx256m -server
On Wed, Sep 15, 2010 at 1:14 PM, Leo Alekseyev dnqu...@gmail.com wrote:
This is a me too post: we just ran into an identical problem setting
up a new cluster using CDH3b2. This is all rather mystifying, because
all the correct libraries are there; in fact the hive command line
looks something
FWIW, the problem in our case was that hive-site.xml enabled map
compression and specified the LZO codec, but we didn't have LZO jar
installed on the new machines. I'm still somewhat puzzled by how the
absence of the LZO library resulted in show tables failing... I guess
the input formats are
Hi, hive-users,
I am a new Hive users, install hive recently, when I type any query-related
command in Hive cli, it throws the exception, but create table are ok:
$ hive
Hive history file=/tmp/pli/hive_job_log_pli_201009141519_1503313446.txt
hive create table test5(a int);
OK
Time taken: 2.551
Some more context: I run Hive on a client machine which is NOT one of the
hadoop cluster nodes, I suppose Hive can run well and submit job from a
client machine, so I didn't change hadoop-env.sh on cluster nodes. On this
client machine, hadoop java jobs and pig have been successfully submitted to