sqoop-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Allaway <andrewalla...@outlook.com>
Subject Re: Sqoop, sending for a loop - newby lost - SQL Server/Sqoop
Date Sat, 02 Nov 2013 01:48:34 GMT
Haaa Devin!  Thankfully I can leave that firewall to the pros. I'm an analytics folk, whose
set up a microCluster at home to learn Hadoop/Hive/Sqoop/db-ETL/cube integration on. And gosh
does it works great now! Best investment in learning I've spent all year - and it's fun to
shut down a node and watch it keep ticking!  

I got the new hardware for a great price, SQL 2014 CTP was free (exp New Year's Eve), Hadoop
free, Hive free, Sqoop free & Debian free!  Gosh is the world changing or what?!  Software
is so open:)

If you or anyone on this thread have or recommend any RDBMS-Hadoop-Hive-Sqoop blogs, websites,
books or whatnot, lemme know.

I'm one'a those who has massive amounts of data to analyze (like many of ya) and am a SAS-SQL
diehard.

Is there any "Kimbal" or "Inmon" of this Hadoop movement I would be sad to not be reading
up on? I read Doug Cutting's blog and follow his tweets and all the big vendors I read up
on (Cloudera, Horton, MapR, Pivotal, etc...).   Anything I'm missing?

I am joining a local Hadoop group in my community (Chicago), so I've the local network part
covered. I've a nice weekend ahead of me and figure now that my cluster is tip-top, it's time
to curl up to a good book, white paper, YouTube/podcast or whatnot.

Best,
Andy



> On Nov 1, 2013, at 8:44 AM, "DSuiter RDX" <dsuiter@rdx.com> wrote:
> 
> Good work! Sometimes it's a simple thing...case sensitivity and the like, or an extra
leading space...glad you were able to track it down.
> 
> Now, if you want another puzzle for later, try putting a firewall in between MSSQL server
and your Hadoop cluster. :-)
> 
> 
> Devin Suiter
> Jr. Data Solutions Software Engineer
> 
> 100 Sandusky Street | 2nd Floor | Pittsburgh, PA 15212
> Google Voice: 412-256-8556 | www.rdx.com
> 
> 
>> On Fri, Nov 1, 2013 at 8:01 AM, Andrew Allaway <andrewallaway@outlook.com>
wrote:
>> Update - 
>> 
>> All's fixed, my TCP/IP port 1433 and UDP 1434 were not open /*blushing*/
>> 
>> Live'n learn:)
>> 
>> Thanks all for help triaging my question!
>> Andy
>> 
>> 
>> 
>>> On Oct 30, 2013, at 10:32 PM, "Andrew Allaway" <andrewallaway@outlook.com>
wrote:
>>> 
>>> Hi Jarcec, et al -
>>> 
>>> Yes, that is rather vague of me.  Apologizes, I appreciate your question and
aid:)
>>> 
>>> I've some updates, but still getting thrown for a loop by Hadoop-Sqoop:)  /*
I can't resist 0>    <--not'a bad ice-cream-cone   :)  */
>>> 
>>> * I was able to uninstall the sqoop 1.4.4 package and re-install it with the
version that says sqoop-bin.1.4.4.  That worked magically!  Woohoo.  I can now run sqoop!
 What was the difference of the stable release with bin vs w/o bin?
>>> 
>>> * Now the down side, I cant, for the life of me, get sqoop to connect to my SQL
Server.  I've the specific jar file to run sqoop for sql server at nameNode: ...sqoop/lib/sqljdbc4.jar
>>> 
>>> Questions:
>>> 1) When I submit the below from .../sqoop/, I get nowhere:)  advice?
>>> 
>>> bin/sqoop list-databases --connect 'jdbc:sqlserver://Andy-ABC-1-HP\BOB:1433;databaseName=andy_dev;user=Andy-ABC-1;password=***;'
>>> 
>>> What I think is the relevant part of error:
>>> ...
>>> 
>>> 13/10/30 06:20:29 ERROR manager.CatalogQueryManager: Failed to list databases
>>> com.microsoft.sqlserver.jdbc.SQLServerException: The TCP/IP connection to the
host Andy-ABC-1-HP, port 1433 has failed. Error: "null. Verify the connection properties.
Make sure that an instance of SQL Server is running on the host and accepting TCP/IP connections
at the port. Make sure that TCP connections to the port are not blocked by a firewall.".
>>>         at com.microsoft.sqlserver.jdbc.SQLServerException.makeFromDriverError(SQLServerException.java:190)
>>>         at com.microsoft.sqlserver.jdbc.SQLServerException.ConvertConnectExceptionToSQLServerException(SQLServerException.java:241)
>>> ..
>>> 
>>> What I've got:
>>> 
>>> a) 
>>> My HP Laptop, set to static: IP4 - 192.168.10.13
>>> SQLServer running on the above laptop:
>>> Host name: Andy-ABC-1-HP
>>> Instance: BOB
>>> Port (per: start>sql server config. mgr>SQL native client>client protocols>tcp/ip>properties>port
1433
>>> DB: andy_dev
>>> schema: dbo
>>> Login: Andy-ABC-1
>>> Pw :
>>> 
>>> b) namenode, 192.168.10.10, Debian, can ping 192.168.10.13 and 192.168.10.13
can ssh into namenode
>>> 
>>> 2) When I lauch sqoop, it keeps saying "Error: /usr/lib/hadoop does not exist!
Please set $HADOOP_COMMON_HOME to the root of your Hadoop installation. & Please set $HADOOP_MAPRED_HOME
to the root of your Hadoop MapReduce installation."  I then run: export export HADOOP_COMMON_HOME=/usr/local/hadoop
and export HADOOP_MAPRED_HOME=/usr/local/hadoop. It works but after reboot it is back??? How
to perm set it?
>>> 
>>> Firewall was off when all above run...
>>> 
>>> Any advice appreciated!
>>> 
>>> Thanks!
>>> Andy
>>> 
>>> > Date: Wed, 30 Oct 2013 14:28:52 -0700
>>> > From: jarcec@apache.org
>>> > To: user@sqoop.apache.org
>>> > Subject: Re: Sqoop, sending for a loop - newby lost - SQL Server/Sqoop
>>> > 
>>> > Hi Andrew,
>>> > would you mind sharing with use exact commands and exact exceptions that
you are seeing? It will help us to understand your issue better.
>>> > 
>>> > Jarcec
>>> > 
>>> > On Mon, Oct 28, 2013 at 07:29:47PM -0500, Andrew Allaway wrote:
>>> > > Hey Abe - 
>>> > > 
>>> > > I haven't because I'm working on running a real lean install on a ARM
cluster.
>>> > > 
>>> > > I want a bare bones build (Hadoop, Hive & Scoop). Starting to feel
that although these packages are open source (Hadoop, Hive, Sqoop, etc), figuring out how
to build a solution from the bottom, w/o a full package (Bigtop, Cloudera, MapR, Horton, et
al), is quite hard, due to not knowing what dependencies are required.
>>> > > 
>>> > > The Doc's of Sqoop 1.4.4 don't mention Hbase being required. What am
I missing?
>>> > > 
>>> > > Anyone tried running a barebones setup and know what I'm missing?
>>> > > 
>>> > > Thanks Abe for the tip, I've your distro on a VM and it has all the
bells and whistles, just was hoping to get a three node solution running swell with the barebones,
just can't figure out what base packages are needed to tie SS 2012/14 and a cluster together
with only Hadoop, Hive and Sqoop.
>>> > > 
>>> > > Also I forgot to mention, when I run sqoop help it also says something
about missing a Java class. I've OpenJDK running, pointing to ARM (i.e. export JAVA_HOME=/usr/lib/jvm/java-7-openjdk-armhf)
>>> > > 
>>> > > Thanks to all in advance for your suggestions!
>>> > > 
>>> > > 
>>> > > Andy
>>> > > 
>>> > > 
>>> > > 
>>> > > 
>>> > > > On Oct 28, 2013, at 6:20 PM, "Abraham Elmahrek" <abe@cloudera.com>
wrote:
>>> > > > 
>>> > > > Andy,
>>> > > > 
>>> > > > Have you tried installing using Apache Bigtop? or some other packaged
installation provider? Hbase client libs are used for Hbase import. Sqoop is compiled with
Hbase support I think.
>>> > > > 
>>> > > > -Abe
>>> > > > 
>>> > > > 
>>> > > >> On Mon, Oct 28, 2013 at 4:04 PM, Andrew Allaway <andrewallaway@outlook.com>
wrote:
>>> > > >> Sorry for the bad title:)
>>> > > >> 
>>> > > >> Have:
>>> > > >> 3 nodes
>>> > > >> Debian/wheezy
>>> > > >> Hadoop 1.2.1
>>> > > >> Hive 0.11.0
>>> > > >> 
>>> > > >> All's working great:)
>>> > > >> 
>>> > > >> Want to connect SQL Server 2012 and SQL Serv. 2014 CTP to
the above
>>> > > >> 
>>> > > >> 
>>> > > >> I'm totally lost
>>> > > >> 
>>> > > >> Namenode (aka node1): 192.168.10.10
>>> > > >> Node2 192.168.10.11
>>> > > >> Node3 192.168.10.12
>>> > > >> 
>>> > > >> Have Windows7 (static ip4 192.168.10.13), connected via ethernet
thru a switch. I can ssh into nodes 1-3 easy.
>>> > > >> 
>>> > > >> All's swell.
>>> > > >> 
>>> > > >> On Win7 have a full sql server instance "bob", database "test_db",
schema "test_schema" & table "test_table" login "abc" pw "xyz".
>>> > > >> 
>>> > > >> On the cluster I've hadoop here:
>>> > > >> /usr/local/hadoop
>>> > > >> 
>>> > > >> Just untared Scoop to /usr/lib/sqoop
>>> > > >> 
>>> > > >> Then when I tried to run$ sqoop help from the above dir, it
said it didn't know where my hadoop was. So I ran the hadoop_home /usr/local....
>>> > > >> 
>>> > > >> Then ran$ sqoop help and it said it can't find hdfs. So I
ran the same$ export home_hdfs usr/local....
>>> > > >> 
>>> > > >> Then ran sqoop help and it said it needs Hbase????
>>> > > >> 
>>> > > >> Does it? Why does it need Hbase to run?
>>> > > >> 
>>> > > >> Not sure how to go from here. I want to install these packages
as I learn them. I don't intend to learn Hbase at the moment, can I "live" w/o it?
>>> > > >> 
>>> > > >> Even if sqoop worked I still don't understand how to pull
the table above (test_table) into hdfs and into Hive??
>>> > > >> 
>>> > > >> Thoughts?
>>> > > >> 
>>> > > >> Best,
>>> > > >> Andy
>>> > > >
> 

Mime
View raw message