Hi all, I am working on setting up Sqoop2 in my CDH4.6 pseudo-distributed environment. Per earlier conversations with Jarek, I'm working toward migrating our usage to #2 vs. Sqoop1 for a few reasons.
I'm coming up with 2 questions. The first is: Is Sqoop2 considered "production-ready?" I only ask because the Sqoop User Guide still seems to be for Sqoop 1. Is the syntax essentially identical when working from CLI, so the guides are essentially interchangeable, and Sqoop2 will provide all the functionality of Sqoop1 but in a server-oriented fashion? The second is - why does my sqoop2 server seem uncooperative? I've yum installed sqoop2-server and client, and uninstalled Sqoop. If I start Sqoop2 server after boot, it tells me a PID file is already there and aborts. Nothing odd there. Stopping it and starting it again yields this output: [root@haddev1 ~]# /sbin/service sqoop2-server start Starting Sqoop Server: [ OK ] Sqoop home directory: /usr/lib/sqoop2 Setting SQOOP_HTTP_PORT: 12000 Setting SQOOP_ADMIN_PORT: 12001 Using CATALINA_OPTS: -Xmx1024m Adding to CATALINA_OPTS: -Dsqoop.http.port=12000 -Dsqoop.admin.port=12001 Using CATALINA_BASE: /usr/lib/sqoop2/sqoop-server-0.20 Using CATALINA_HOME: /usr/lib/bigtop-tomcat Using CATALINA_TMPDIR: /var/tmp/sqoop2 Using JRE_HOME: /usr/java/jdk1.6.0_31 Using CLASSPATH: /usr/lib/bigtop-tomcat/bin/bootstrap.jar Using CATALINA_PID: /var/run/sqoop2/sqoop-server-sqoop2.pid >From there, I attempt to test the server response per the Cloudera guide with the command: wget -qO - localhost:12000/sqoop/version but I don't get the JSON token I'm supposed to get back. If I enter the client and request version info, I get this: sqoop:000> show version --all client version: Sqoop 1.99.2-cdh4.6.0 revision Compiled by jenkins on Wed Feb 26 02:44:40 PST 2014 Exception has occurred during processing command Exception: com.sun.jersey.api.client.UniformInterfaceException Message: GET http://localhost:12000/sqoop/version returned a response status of 404 Not Found So, I'm kind of stumped at that point. Is there some contraindication regarding running Sq2 client on the same box as the server, essentially stopping it from being worked with in pseudo-distributed mode? Or does something stand out as a problem with my config/is there some debug I can pull out to glean more insight into why the service isn't listening on that port, even though it appears to be? Thanks for any assistance you can give. I'd like to avoid spinning up a fully-distributed VM cluster to test Sqoop2 jobs on, if possible, and my real cluster is indisposed at the moment... *Devin Suiter* Jr. Data Solutions Software Engineer 100 Sandusky Street | 2nd Floor | Pittsburgh, PA 15212 Google Voice: 412-256-8556 | www.rdx.com
