For reference, I've created OOZIE-2310 to improve this situation. On Fri, Jul 17, 2015 at 11:39 AM, Robert Kanter <[email protected]> wrote:
> Also, did you setup Oozie's Hadoop settings? > http://oozie.apache.org/docs/4.2.0/AG_HadoopConfiguration.html > The oozie.service.HadoopAccessorService.hadoop.configurations property is > very important. > Make sure to do that before the sharelib command I mentioned. > > On Fri, Jul 17, 2015 at 11:34 AM, Robert Kanter <[email protected]> > wrote: > >> Hi Matteo, >> >> I took a quick look at the code and it looks like you got that ugly (and >> not helpful NullPointerException) because of OOZIE-1877. I'll create a >> JIRA to fix that. >> >> Anyway, to fix the problem, you need to deploy the ShareLib in HDFS. >> oozie-setup.sh sharelib create -fs hdfs://HOST:8020 >> >> >> - Robert >> >> On Fri, Jul 17, 2015 at 1:06 AM, Matteo Luzzi <[email protected]> >> wrote: >> >>> Hello! I'm new on this mailing list since I just started looking into >>> oozie >>> >>> I'm trying to run oozie 4.2.0 on hadoop 2.7.0 working in a >>> psudo-distributed mode on my local machine. I built a distro of oozie >>> binding it to the correct hadoop version I'm using, then, according to >>> the >>> documentation I did the following steps >>> Created a folder called *libext* where I put a .zip of *ext-2.2* library >>> Lauched the commands : oozie-setup.sh prepare-war, ooziedb.sh create >>> -sqlfile oozie.sql -run, oozied.sh start >>> >>> Everything went fine, in fact I can navigate localhost:11000/oozie/ and >>> also if I execute oozie admin -oozie http://localhost:11000/oozie >>> -status the >>> server replies with *NORMAL*. >>> >>> However I'm not able to launch any jobs, neither the examples. After >>> moving >>> the example folder in the correct position on HDSF if I execute >>> oozie job -oozie localhost:11000/oozie -config >>> /path/to/examples/apps/map-reduce/job.properties -run the server replies >>> with >>> *Error: HTTP error code: 500 : Internal Server Error*. >>> Same situation if I try to submit a workflow programmatically using the >>> Java API. >>> >>> Looking at the oozie.log file I get the following error/warning: >>> >>> ERROR V2AdminServlet:517 - SERVER[myserver] USER[-] GROUP[-] TOKEN[-] >>> APP[-] JOB[-] ACTION[-] URL[GET >>> http://localhost:11000/oozie/v2/admin/instrumentation?_dc=1437056225186] >>> error, >>> >>> null java.lang.NullPointerException at >>> >>> org.apache.oozie.service.ShareLibService.getLatestLibPath(ShareLibService.java:687) >>> at >>> >>> org.apache.oozie.service.ShareLibService$7.getValue(ShareLibService.java:742) >>> at >>> >>> org.apache.oozie.service.ShareLibService$7.getValue(ShareLibService.java:737) >>> at >>> >>> org.apache.oozie.servlet.BaseAdminServlet.instrElementsToJson(BaseAdminServlet.java:312) >>> at >>> >>> org.apache.oozie.servlet.BaseAdminServlet.instrToJson(BaseAdminServlet.java:339) >>> at >>> >>> org.apache.oozie.servlet.BaseAdminServlet.sendInstrumentationResponse(BaseAdminServlet.java:396) >>> at >>> >>> org.apache.oozie.servlet.V2AdminServlet.sendInstrumentationResponse(V2AdminServlet.java:124) >>> at >>> >>> org.apache.oozie.servlet.BaseAdminServlet.doGet(BaseAdminServlet.java:127) >>> at javax.servlet.http.HttpServlet.service(HttpServlet.java:617) >>> >>> everytime I refresh the webapp interface and >>> >>> AuthenticationToken >>> ignored:org.apache.hadoop.security.authentication.util.SignerException: >>> Invalid signature >>> >>> everytime I submit a workflow either via command line >>> >>> I cant even get it working in local mode. I get all the jobs sumbitted to >>> local server killed. I thought it could have been an hadoop problem, but >>> everything seems fine >>> >>> My working environment: >>> >>> Mac Os Yosemite >>> java version "1.8.0_45" >>> Hadoop 2.7.0 >>> oozie 4.2.0 >>> Apache Maven 3.3.3 >>> >>> Can anyone help me to get it working? >>> >>> -- >>> Matteo Remo Luzzi >>> >> >> >
