Hmm. The data in my tables is not important. So, I dropped the table and
recreated it. This doesn't seem to have resolved the issue, though.
Is there perhaps a Pig query I can run that would use a built in HBase table,
like that .META. table, and see if it works? I don't know if that'd help or
not, though.
Ryan
On Mar 22, 2012, at 10:02 PM, Norbert Burger wrote:
> Actually on second glance, this seems like an issue not with the HBase
> config, but with the server:port info inside your .META. table. Have you
> tried LOADing from a different table besides "events"? From the HBase
> shell, you can use the following command to extract server hostnames for
> each of your regions:
>
> scan '.META.',{COLUMNS => 'info:server'}
>
> Dropping and recreating the table should resolve the issue; if you have
> real data already, then there are tools in hbase/bin to repair regions.
>
> Norbert
>
> On Thu, Mar 22, 2012 at 10:44 PM, Ryan Cole <[email protected]> wrote:
>
>> I was thinking that maybe it was because I did not have HBase config path
>> on PIG_CLASSPATH, so I added it. This did not help, though.
>>
>> Ryan
>>
>> On Thu, Mar 22, 2012 at 9:07 PM, Ryan Cole <[email protected]> wrote:
>>
>>> Norbert,
>>>
>>> I have confirmed that this is indeed an issue connecting to HBase. I
>> tried
>>> just running a Pig script that did not use HBaseStorage, and it works.
>> Here
>>> is my hbase-site.xml config file, as well as my query that I'm running:
>>>
>>> https://gist.github.com/2166187
>>>
>>> Also, for ease of reference, here is the query:
>>>
>>> raw = LOAD 'hbase://events' USING
>>> org.apache.pig.backend.hadoop.hbase.HBaseStorage('event:*', '-loadKey
>>> true') AS (id:bytearray, events_map:map[]);
>>>
>>> Maybe I need to change the rootdir in the config to be my hostname, and
>>> not localhost?
>>>
>>> Ryan
>>>
>>>
>>> On Thu, Mar 22, 2012 at 8:58 PM, Norbert Burger <
>> [email protected]>wrote:
>>>
>>>> You're encountering problems connecting to HBase (presumably your Pig
>>>> script uses HBaseStorage). How does your hbase/conf/hbase-site.xml
>> look?
>>>>
>>>> Norbert
>>>>
>>>> On Thu, Mar 22, 2012 at 9:16 PM, Ryan Cole <[email protected]> wrote:
>>>>
>>>>> Hello,
>>>>>
>>>>> I'm new to these lists. I'm trying to get Pig working, for my first
>>>> time. I
>>>>> have setup Hadoop and HBase (on HDFS) using the psuedo-distributed
>>>> setup,
>>>>> all on one machine. I am able to run MapReduce jobs, using the
>>>> example.jar
>>>>> file included with the Hadoop release.
>>>>>
>>>>> Whenever I try to run even the simplest query examples, using Pig, I
>> get
>>>>> the following error:
>>>>>
>>>>> `ERROR 2017: Internal error creating job configuration.`
>>>>>
>>>>> and, the log file has the following more specific error:
>>>>>
>>>>> `Caused by: java.lang.IllegalArgumentException: Not a host:port pair:
>>>>> �^@^@^@^P8948@ryan-serverlocalhost,44544,1332443083936`
>>>>>
>>>>> It looks like Pig goes through the entire compile process, of turning
>>>> the
>>>>> Pig Latin into the MapReduce code, but fails to send it off to
>> Hadoop's
>>>>> MapReduce. That's just my uneducated analysis of what I see happening,
>>>>> though. I have pasted the Grunt console output, and the log file
>>>> contents
>>>>> here: https://gist.github.com/2163762.
>>>>>
>>>>> This is Pig v.0.9.2, HBase v. 0.92.1, Hadoop v1.0.1.
>>>>>
>>>>> Does anyone have any idea why this may be happening? It looks obvious
>>>> that
>>>>> I have something configured improperly, but I looked for any host:post
>>>>> settings that would stand out and didn't find anything obvious.
>>>>>
>>>>> Thanks,
>>>>> Ryan
>>>>>
>>>>
>>>
>>>
>>
