Hi Johannes,

Currently, the platform does require the republish of the applications after an upgrade. Since O10, the whole installation process can automated using the Unattended Installation API. You can check it here.  

The OSPTool has a command that allow you to republish the entire factory without the need of create an solution first:

• outsystems_common_path\OSPTool.com /PublishFactory hostname username password

This doesn't solve possible problems with broken dependencies between the versions but allow you to automate the upgrade process.

Does this address your concerns?

I don't think that telling people to write a bunch of code and somehow test it to get an unattended installation is what people asked for. That's a nice thing to have in a large very enterprise/DevOps environment, but that does NOT describe most current customers. Especially since so many on-prem customers only upgrade a few times a year. The "cost" to automate this process is far higher than the cost of doing it manually.

If there is an idea for "we want an API for an unattended installation/upgrade" that should get marked as "implemented",  but what people are asking for here was not a method to automate upgrades, but a way to make manual upgrades easier.

J.Ja

Hi John,

Thank you for your contribution. We merged your idea with a previous one that already existed with the same purpose.

Regarding your comment in concrete, I would like to ask you a couple of questions:

1. You mentioned that during the republishing of your factory, a publication that evolves all content (around 1000 eSpaces), you end up on getting Internal Errors. Can you elaborate what kind of error or message does it show?
2. Currently, there is a way to trigger that operation (republish the whole factory) via OSPTool. You can find the documentation of that API here.  Let me speculate a bit: imagine that after an update when the installation of the Service Center ended, you were prompted if you want to republish all factory for that environment? (something that you need to do since it's a mandatory step from the installation checklist). Do you think that having it more visible and actionable would help?

Cheers!

Hi Tiago, 

For point 1, That was the exact error message, just "Internal Error" nothing else. Support helped me with the actual issue in the end, there was a timeout during the addition of all espaces, most likely due to the number of espaces and the size of the espace version entities.

For Point 2, yes that would help, however what would be far more useful is for the dev environments to just automatically proceed with the republish as part of the platform upgrade. During this upgrade process it took me over 15 hours (no I'm not exaggerating, included having to restart half way due to the IIS app pool being recycled by support troubleshooting the first issue)

Because the republish is mandatory it would save lot of time and outage if support could just trigger the process for us after they finish the base environment upgrade.   Not for Production though, I agree that we need a lot more control and transparency for that environment.

Thanks for your insights John.

Just to clarify, in this particular case, you are talking about an environment hosted in our OutSystems Cloud right?

Yes this is a cloud hosted system.