1 of 2
1
New FAST build
Posted: 28 October 2009 10:51 AM   [ Ignore ]
Member
RankRankRank
Total Posts:  54
Joined  2009-01-22

I saw the blog post about the new FAST build but I don’t see anything new in Github. I’d like to pull the latest release so I can merge with ours.

Profile
 
 
Posted: 29 October 2009 03:52 AM   [ Ignore ]   [ # 1 ]
Administrator
RankRankRankRank
Total Posts:  249
Joined  2008-12-18

It should be available later this morn. We’re trying to find the best way to export the project with commit history. Stay tuned.

Profile
 
 
Posted: 29 October 2009 10:42 AM   [ Ignore ]   [ # 2 ]
Member
RankRankRank
Total Posts:  54
Joined  2009-01-22

I See int there now…

Thanks

Profile
 
 
Posted: 29 October 2009 10:46 AM   [ Ignore ]   [ # 3 ]
Administrator
RankRankRankRank
Total Posts:  249
Joined  2008-12-18

Well it’s just an empty project until we do the import. We’re running into problems maintaining the commit history. Bare with us and we may have to ignore it for this go-round.

Profile
 
 
Posted: 29 October 2009 10:56 AM   [ Ignore ]   [ # 4 ]
Member
RankRankRank
Total Posts:  54
Joined  2009-01-22

Works for me.. I don’t really look at the history!

Profile
 
 
Posted: 29 October 2009 11:45 AM   [ Ignore ]   [ # 5 ]
Administrator
RankRankRankRank
Total Posts:  249
Joined  2008-12-18

We’ll have to save the commit history for the next release. Our git-svn conversion didn’t go as well as planned, but at least the latest code is now available:

http://github.com/numerex/FAST_CORE_Tracking

Profile
 
 
Posted: 03 November 2009 03:11 AM   [ Ignore ]   [ # 6 ]
Member
RankRankRank
Total Posts:  54
Joined  2009-01-22

I’ve merged 10340 with our build, including all the revised deployment files. Over the past few weeks there seems to have been some back and forth regarding the correct process - deploying from trunk, deploying from a tag etc. Now that you have hopefully got the process to the way you want it, can you do a blog post or something on a definitive guide to deploying using the latest build?

Profile
 
 
Posted: 03 November 2009 08:52 AM   [ Ignore ]   [ # 7 ]
Administrator
RankRankRankRank
Total Posts:  249
Joined  2008-12-18

Glad to hear you got it merged. Yes we have the process nailed down pretty good and will definitely blog it. Great suggestion!

Profile
 
 
Posted: 04 November 2009 02:59 AM   [ Ignore ]   [ # 8 ]
Newbie
Rank
Total Posts:  6
Joined  2009-10-14

can you do a blog post or something on a definitive guide to deploying using the latest build?

Voting for this one too! Would appreciate it very much.

Profile
 
 
Posted: 20 June 2010 08:57 PM   [ Ignore ]   [ # 9 ]
Newbie
Avatar
Rank
Total Posts:  26
Joined  2009-02-22

Hey guys - just curious to know if there was ever any post with a full build process described. I checked the actual blog, but haven’t seen one.

From what I understand, the general process should be:

1. Numerex commits code changes, etc. to github
2. Customers (like me) grab the latest files from github and merge them into their own repository (trunk).
3. Customers test locally to ensure all is well.
4. Customers deploy and test on staging (cap staging deploy)
5. Customer deploy and test on production (cap prod deploy)

The questions I have revolve around your suggested best practices for SVN labeling, etc. I see in our repository several tags labeled “successful_build_xxxx”, “current_staging_release” and “current_prodution_release”.

I’m assuming its OK to do my merges from github into my trunk and just do the cap deployments right from there. Then, apply labels as needed. However, I don’t know the cap process very well - does it pull from the trunk or from specific tags?

Thanks
Scott Leonard

Profile
 
 
Posted: 21 June 2010 01:10 PM   [ Ignore ]   [ # 10 ]
Administrator
RankRankRankRank
Total Posts:  249
Joined  2008-12-18

Scott,

The successful_build directories are managed by our automated build system (aka continuous integration). Unless you specifically ask to be included in this system you should ignore the successful_build directories and deploying from the trunk directly. The current_staging_release and current_production_release tags are used when tagging a successful build from the CI system. To circumvent the need for using our CI system we’ve provided a “DIY” deployment file that you can see here:

http://github.com/numerex/FAST_CORE_Tracking/blob/master/config/deploy_diy.rb

You just need to replace the following strings with your appropriate info:

YOUR_SVN_USERNAME
YOUR_SVN_PASSWORD
YOUR_PRODUCTION_SERVER_ADDRESS
YOUR_SVN_BRANCH_NAME
YOUR_STAGING_SERVER_ADDRESS

I hope this helps and let us know if you have any more questions.

Profile
 
 
Posted: 22 June 2010 07:34 PM   [ Ignore ]   [ # 11 ]
Newbie
Avatar
Rank
Total Posts:  26
Joined  2009-02-22

Thanks Dennis - this helped. However, when I run the build to our staging environment the web app errors out immediately with a “Bad Gateway” message.

The cap deployment seemed to run fine.

This is using the latest files from github. We have not made any code changes ourselves since the last build.

Any suggestions?

Thanks,
Scott

Profile
 
 
Posted: 22 June 2010 07:47 PM   [ Ignore ]   [ # 12 ]
Administrator
RankRankRankRank
Total Posts:  249
Joined  2008-12-18

There may be pending migrations that need to be run since your last deployment. Did you do a cap staging deploy:migrations?

Profile
 
 
Posted: 22 June 2010 09:25 PM   [ Ignore ]   [ # 13 ]
Newbie
Avatar
Rank
Total Posts:  26
Joined  2009-02-22

Thanks for the tip - just tried it and am still getting the same result.

Profile
 
 
Posted: 24 June 2010 11:22 AM   [ Ignore ]   [ # 14 ]
Administrator
RankRankRankRank
Total Posts:  249
Joined  2008-12-18

We just took a look at your staging system and for reference you should always check in the mongrel logs when you get a bad gateway error. This was in the log:

/usr/lib/ruby/gems/1.8/gems/activesupport-2.3.2/lib/active_support/dependencies.rb:105:in `const_missing': uninitialized constant IdleEvent::SuspectEvent (NameError)
        from /opt/ublip/rails/releases/20100623031055/app/models/idle_event.rb:2 

After looking at your deployment it appears that you don’t have the latest build from GitHub. In this example it’s looking for the SuspectEvent model, which doesn’t exist on your server.

Profile
 
 
Posted: 21 September 2010 12:00 PM   [ Ignore ]   [ # 15 ]
Newbie
Rank
Total Posts:  2
Joined  2010-09-16

hey guys i also was just wondering if you had full build process described anywhere. that would be helpful.
thanks.

Profile
 
 
   
1 of 2
1
 
‹‹ GPIO Data/Events      Job Scheduling ››