Ask Rallypublic

Questions & Answers

Get your Rally questions answered.

This is a public Hive  publicRSS

Question

    Rally Integration with Subversion
    Question posted 07/31/09 by J-P
    8377 Views, 5 Comments
    Title:
    Rally Integration with Subversion
    Summary:

    Hi All!

    We are considering Rally adoption. At this stage, we are still collecting and analyzing information relevant to various aspects of the development life cycle. We are now currently investigating Rally integration with Subversion (management just asked me tolook into this matter). Thus here are a few questions:

    (1). What is involved in the integration process?

    (2). How much effort (realistically) can be involved?

    (3). Where can I find resources about the Integration process?

    Any helpful tip will be highly appreciated.

    J-P

    Comments

    • posted 08/07/09 by Dave Smith

      Hi JP - the integration runs as a ruby script you plug into your post commit hook.  

      There is a readme in the zip file which you can get from the subversion post here which goes over the general setup.  You will need to modify the script in a few places to setup which user and workspace the connector will use to update Rally along with the necessary setup steps to get the Ruby script running properly.  

      Overall, if you are comfortable with installing Ruby and managing Subversion (eg installing the hook), the install and test should not take more than a few hours - if everything goes quickly (or you have a playgrounf for svn), it should not take more than an hour to at least get a proof of concept to test.

      Reply to this Comment

      Attach File...

    • posted Feb 03 by Kapil

      Hi,

      I followed instructions to connect my local test SVN server (running on Snow Leopard MacBook Pro) with Rally.

         1. Skipped ruby install since it comes preinstalled on OS
         2. Didnt find any "CachedStates" node within root element, so added one with default values posted in instructions.
         3. Confirmed config file was correct by running "ruby config_test.rb"
         4. Next, updated svn > hooks > post-commit (copy of post-commit.tmpl) with reference to connector

      Did a dummy checkin via eclipse with comments "TA3 Completed". Dont see any updates to my task.

      Rally_Connector_Error.log reports following - which doesnt reveal much info. I removed spaces in workspace name later - no luck.

      # Logfile created on Wed Feb 03 15:10:38 -0700 2010 by /

      E, [2010-02-03T15:10:38.655984 #3992] ERROR -- : RALLY_WORKSPACE_NAME - Info Trust Group Product Engineering

      Anybody got any tips how to troubleshoot it further?

      Thanks
      Kapil

      Reply to this Comment

      Attach File...

    • posted Feb 04 by Charles Ferentchak

      There is a chance that your cached states are off in your file. The first time you run the connector it will cache those in the XML file for you. You can try deleting that entire element from your config file and running everything again to make sure there is not a weird character or a space somewhere there shouldn't be. 

      You could add this to your config file to get more logging info too.

      <Log>
<FileName>
 test.log 
</FileName>
<Level>
 0 
</Level>
</Log>


      


      

      Let me know if that helps.

      -Charles

      Reply to this Comment

      Attach File...

    • posted Feb 03 by Kapil

      Charles, thanks for prompt response. I enabled log and it is reporting following error if I add "cachedstates" manually to config.xml

      I, [2010-02-03T16:41:28.400975 #1328]  INFO -- : RALLY_USERNAME - ktundwal@infotrustgroup.com
      I, [2010-02-03T16:41:28.402242 #1328]  INFO -- : RALLY_BASE_URL - https://rally1.rallydev.com/slm
      I, [2010-02-03T16:41:28.402831 #1328]  INFO -- : RALLY_WORKSPACE_NAME - InfoTrust_Group_Product_Engineering
      E, [2010-02-03T16:41:28.405554 #1328] ERROR -- : RALLY_WORKSPACE_NAME - InfoTrust_Group_Product_Engineering

      Checking from eclipse doesnt generate any further logs.

      Reply to this Comment

      Attach File...

      • posted Feb 04 by Charles Ferentchak

        When I looked into this it appears that the caching on the states has some errors in it that we will be addressing in the next version. In my testing the code still functions but is a tad slower then I would like and will throw the error message you are seeing even if the connector is set up correctly. In short I would guess that that error is not what is keeping your connector from working and it may be best to go down a different path to debug.

        -Charles

        Reply to this Comment

        Attach File...