Open HeWessel opened 1 year ago
Hello we face the same issue, is there any Solution or Workaround to solve ?
Was the app installed/upgraded? I am asking in order to learn whether the problem is during getting the install/upgrade status - or whether the problem actually occurs when trying to install/upgrade
We recently also have issues when publishing from vscode. It seems like the sandboxes are generally less reliable then they used to be. Somethings publishing an app takes 10 min.
Until now we only faced the Issue during upgrade the App. I will try to reproduce the Problem while publishing a new App to the Sandbox.
We face the issue using Devops Release Pipeline with outbound Powershell.
Does this problem still occur? Does anybody have a good repro?
Hi Freddy, yes the behaviour is still there. We currently have an workaround doing a Try and Catch, wait 3 minutes and then retry to install the app. Usually after 3 minutes the app is installed and when then second run is triggered it will say "app already installed" and the rollout is succesfully. How can i help you find the issue ? It could be repo installing an Devops Agent on your local PC an Deploy to an BC Online.
Thanks - will investigate and get back
Hi Freddy below the error Message still showing up
"appname_21.0.0.1.app - publishing and installing...ERROR The server committed a protocol violation. Section=ResponseStatusLine
at
Hey Freddy,
i was able to check the different circumstances. Here is what I noticed so far. Publish in a fresh Environment
I also saw one time that the App Upgrade was started at the same time to the Production and the Sandbox Environment. The Upgrade in the Sandbox failed and the Upgrade in the Production had no Problems. A few days later it was the opposite.
To temporary handle the Problem we changed our Pipeline and set the "retryCountOnTaskFailure" in the Pipeline. If any Error occurs while Upload the App, a Delay will start, before the next Retry. So the Script will try again after a few Minutes and see that the right Version is already installed. But thats hopefully only a temporary Solution for the Problem.
I hope that can help you. If you need any further informations let me now.
Thanks
Would it be possible to get the apps you are publishing to see whether I can repro using your script above?
if i remember correctly it occurs randomly and only with a local Agent. When using an Azure Agent there is no problem.
I also made tests with the Al Go Project, where the Costumerlist is only extended with an OnOpenPage Trigger. And I had the same Results. I attached the Example Project. But we also only uses local Agents.
For AL-Go projects, you could try to set:
"runs-on": "ubuntu-latest",
"gitHubRunner": "self-hosted",
"gitHubRunnderShell"; "powershell"
in .github/AL-Go-Settings.json
and run Update AL-Go System Files.
This would cause you to use self-hosted runners for the time consuming build job, but GitHub hosted Linux runners (cheap and fast) for the rest.
For AL-Go projects, you could try to set:
"runs-on": "ubuntu-latest", "gitHubRunner": "self-hosted", "gitHubRunnderShell"; "powershell"
in .github/AL-Go-Settings.json
and run Update AL-Go System Files.
This would cause you to use self-hosted runners for the time consuming build job, but GitHub hosted Linux runners (cheap and fast) for the rest.
Hi Freddy do you have some setting suggestions for Devops ?
@Pomaroli - that all depends on the solution/script you are using on DevOps. AL-Go has a lot of functionality, which is controlled by settings - this is not GitHub or Azure DevOps functionality.
Okey i understand, we not using AL-Go for now maybe in future. We are using Azure Devops and i developed myself Build and Release Pipelines based on YAML and Powershell. To Compile and Test the AL-Source code we use Docker Instance's. After the App is compiled sucessfully we deploy using the In-Build Release-Pipeline functionality of Azure Devops with Powershell. Basically we authenticate against BC Online an then Deploy PTE nothing special there. Maybe im completely wrong but i could imagine the issue is caused because of timeouts on responses occuring while the System ist upgrading.
@freddydk sorry for the late response. I forgot that there is also the Al-Go Project on GitHub. I just wanted to tell you that I just created in VS-Code a basic project with the function AL-GO. So there is an app initiated with the Page Extension on CustomerListExt with the "Hello World" Messagebox. I had the same problem uploading this basic app with one Page extension to my Saas Environment. For the moment, I extended our pipelines so that the PTE-Upload retries after a delay of 5 minutes. Afterward, the app is published, and no app will be uploaded again. So at the moment, I tried the first install and also upgraded over self hosted Agents. The problem still sadly occurs. Any other suggestions on what I can try? If you need any further information let me know.
@freddydk This problem still bothers us. Especially when we try to deliver more than one App into the Cloud and the Problem occurs at every App. Any Updates on this topic?
@HeWessel - did you try to switch to GitHub or Azure DevOps hosted runners to see whether this solves the issue? I will grab the sample you provided last and make some tests with that to see if I can repro.
Hi There, @freddydk did you changes something ? Since some days it seems do work fine now. Anybody else noticed some new behaviour ?
2024-01-10T09:26:38.7571725Z xy.app - upgrading....................completed
I didn't - maybe they fixed something in BC online to solve this
It did it again, sometimes it doesn't work. I haven't figured out what the connection is yet
Describe the issue
We are using the function Publish-PerTenantExtensionApps in our Scripts to automate Publish new App versions to SaaS Environments. There are recently some Problems with the result of the Function. The strange thing is, that sometimes we can publish without any Problems and sometimes the Error occurs. In both Cases the App will be uploaded and also installed. Are we missing any Params for the function or is there a known issue of my problem? The Problem occurs also on different Environments but that can be a result of sharing the same Powershell Library and using the same script with different Params for our function.
Scripts used to create container and cause the issue
Full output of scripts
Additional context