Deployment and upgrades - progress and challenges

Hi all

We are excited to be planning for a couple of months with three developers in the room here in Aus, and keen to get as much stuff sorted in this time as we can.

One area we want to try and include is streamlining and supporting deployments. If you are or have been working on deployments, upgrades etc can you please reply to this post with a quick note to let us know:

  • what you’re trying to do;
  • where you’re up to;
  • if/where you are stuck (linking to other discourse posts or git issues where relevant), what you’ve tried and any progress on narrowing things down, and
  • any relevant timelines e.g. trying to get pilot launched for XXX by XXX

Please also let us know if you’re wanting to upgrade to next version but don’t have development capacity to do so

@NickWeir @lin_d_hop @pmackay - s’ok we’re across UK, you don’t need to respond

@sigmundpetersen @gnollet @MikeiLL @lawrence @sylvain @MyriamBoure

please ping anyone I may have missed . .

Yeahhhhhhh, what a great news @Kirsten ! I love this message :smile:

For the French deployment, I let @gnollet answer as he has been the most involved, along with @marito59 in trying to deploy the platform. I now we are stuck :-o
We have forgotten the idea to have that first pilot on the 26th September, so we don’t have a formal date now, but the sooner the better :slight_smile:

For the Norwegian deployment, we definitely want to upgrade. @sigmundpetersen said he is keen to try to upgrade the server but he saw your message saying you were working on an upgrade procedure and is waiting for this support to move forward. He also thought that coudl make sense to wait for the language framework to be merged, so that we can test both the design customization feature and the front-end translation (ping @maikel).

Ping @CynthiaReynolds for information, @Selmo if you know about other instances in deployment process.

Hi. I too am excited by this message. Over here in the states, as far as deployment, I got stuck at Need help to fix ERR 500, using the Vagrant (ofn_deployment) install. There also seemed to be issues with seeding the suburbs, which for USA is a very large file, but from what I read this may not matter much at the moment. It’s likely that my error is due to some obvious step that an experienced Rails developer would have done like a rake db:migrate or something.

Have been focusing my energies on getting up to speed on Ruby and Rails and am looking forward to chatting with Anselm at some point soon, and potentially connecting with some of the other interested parties in the USA that Myriam mentioned.

It would be very nice to get a working development deployment as well as staging if not production up as soon as possible as our current php system has never been terribly efficient or even reliable. : )

Hi everyone.

Here in South Africa we are about to launch. We merged with the master branch this week so are up to date with the latest code. Apart from some mailing issues which we eventually tracked to a malformed URL, it has been a relatively painless journey. We opened for pre-registration today and will be hosting two launch events on Monday 21’st and Tuesday 22’nd in Johannesburg and Cape Town respectively.

Some snags we encountered and are dealing with right now:

* Superuser can't delete users from the admin panel - a grumpy cat jumps out
* Map centres on Australia - we want it to zero in on user's location
* Some minor local configuration issues

I’ll update everyone after the launch

Have a good weekend!

Lawrence

1 Like

@gnollet Do you have anything to add about France or is it the same as @MikeiLL?

I think we have the same issue than @MikeiLL

@Kirsten we actually have another problem in Norway. We only have a production environment, so we do all our tests in it, and it starts to be very messy but we want to launch the first real life experimentations. In the deployment process you helped us with, Raf only set up a production environment. But it can be very risky as we start communicating about some test, and if people start to go around and order in 'fake" order cycles, that may be a problem… I don’t know how to solve that… maybe @sigmundpetersen you will have an idea?

Yes we should have test and staging servers and follow the architecture as in https://en.wikipedia.org/wiki/Deployment_environment. Thus we can have stable releases running on the production server and test new versions on test and staging. But we need to get more experienced in setting up the servers or get people on board that are experienced in doing this :smile:

Hi,

I will be pleased to help you on this part as soon as the ERR 500 issue will be fixed.
Then I wille be able to test and deploy version on required envirronement.

1 Like

I’ve been keeping a fairly detailed log of deployment efforts and wanted to encourage others to do the same in a publicly available space so that we can more easily learn from each others challenges and how they are overcome.