none
503 Site Under Construbtion after GitHub deployment shows Success

    Question

  • I have had zero issues with continuous deployment for about two years now on two deployment slots for an Azure Web App. Suddenly, (meaning 11PM Pacific time on Feb 10, 2019) one of the deployments results in a 503 server error while showing "Success (Active)" under the Deployment Center. The logs also show no errors. I have tried resyncing, disconnecting and reconnecting the repo. I have even tried using another branch, and the result is always a 503.

    My other Web app slot setup doesn't have this problem. Both are slots on the same App Service and the logs show the exact same results. I'm at a loss for figuring out why one slot would work but another slot on the same App Service wouldn't work.

    Anyone else have a similar experience?

    Monday, February 11, 2019 7:50 AM

All replies

  • Thanks for reporting the issue! Just to clarify, do you recall making any specific changes prior to this issue? On which region have you provisioned your app?

    1.Kindly check to see if App Service diagnostics provides any pointers, it reduces the chance of trial and error and expedites problem resolution by recommending potential solutions.

    To access App Service diagnostics, navigate to your App Service app in the Azure portal. In the left navigation, click on Diagnose and solve problems.

    2.You could also review the Service Health notifications.

    The ‘Service Health’ -Service issues view shows any ongoing problems in Azure services that are impacting your resources. You can understand when the issue began, and what services and regions are impacted. You can also read the most recent update to understand what Azure is doing to resolve the issue. See, Azure Service Health

    If the issue still persists, could you share your WebApp name directly here or indirectly?

    Monday, February 11, 2019 1:33 PM
    Moderator
  • Thanks for the feedback. There were not changes made on the Azure side. The repo was update, and it auto deployed to one slot just fine. The same repo was then merged into the branch used for this slot that has problems and the result is the 503 on that slot. Same code, two different slots. One is fine; One is 503.

    On the Azure portal, there is no Diagnose and solve problems in the left nav. Under Diagnostic Settings only the Azure Database resources are showing as options, and there is no option under the Resource type dropdown for Web App.

    I made a new slot and set it up to pull from the same repo branch that's having this 503 error, and the WebApp is running just fine on that slot. WebApps are em…-auto (returning 503) and em…-a2 (works fine).

    At this point since the same repo branch is working fine on a newly created slot, I'm not too worried about this issue. But I'm definitely interested in discovering what might have gone awry with the WebApp that's having the problem while other slots are not having any issues.

    Thanks again!

    Monday, February 11, 2019 6:16 PM
  • Try checking d:\home\logFiles\eventLog.xml see if it yields any clue.    The one stood out seemed to be "Unsecured Passwords Format Detected. The Membership Provider that contains the unsecure passwords format is: MySQLMembershipProvider. The obsoleted password format is: Clear. For more information, see https://go.microsoft.com/fwlink/?linkid=834784."   Are you using that by any chance or connection string messed up?

    Suwatch

    Monday, February 11, 2019 6:52 PM
  • That's very odd. There is not anywhere in any file MySQLMembershipProvider in the project, let along in web.config. When streaming the Application Log, I'm seeing the following:

    Physical Path   D:\home\site\wwwroot Logon Method   Not yet determined Logon User   Not yet determined

    This is very odd. It's even more bizarre since the same branch deploys without error to a new slot that I created. No changes to anything, just a new slot that didn't copy the WebApp settings from the slot that's having problems. I also made another slot with the settings copied from the slot that's going 503, and that slot is working fine using the same exact repo branch. Not sure what happened with this particular WebApp that's gone 503.

    Tuesday, February 12, 2019 12:06 AM
  • You can also check to see if there is an appoffline.html file in your wwwroot directory. If so this may have been added during the build and not properly cleaned up.
    Tuesday, February 12, 2019 2:09 AM
  • Thanks. We decided to just delete the slot and then add it back. Everything is working fine now.
    Tuesday, February 12, 2019 4:30 AM