Azure pre-configured VM - fails to load after changing config

Hello,

I’ve been running Wowza instances in Azure for about a year now (custom built) and decided to try the recently introduced pre-configured VMs. Initially everything went fine following the provided guide.

However, once I make changes via the Management Console, or change my own custom code, and restart Wowza, the system doesn’t return to operational. The management page remains up, but it is unable to contact the underlying API. Attempting to reach the server at port 1935 shows that Wowza is down/unresponsive.

I am able to verify via SSH that both the Engine & Manager are running. I do not see any errors logged (in fact the normal log files are not even touched) so it appears the issue occurs prior to my code being loaded.

Rebooting the VM itself does not fix the problem either.

Any help you can offer on how to debug or find helpful startup logs would be great.

Thanks

Chris

Here is the exact error from the Management Console:

Wowza Streaming Engine Manager could not connect to the Wowza Steaming Engine (http://localhost:8087). Verify that the Wowza Streaming Engine service has started and is running.

I have verified that the correct ports are open for the VM (default from the pre-configured package)

Hi,

It does suggest a bug in the Azure infrastructure. If the reboot in the command line works, and assuming that is what the DashBoard request tries to do behind the scenes then it would be something to open a support request with Azure.

As you have checked Wowza is running as it should be then it suggests a network/firewall set up problem after the reboot in the Azure DashBoard infrastructure.

Andrew.

This issue has been received in support ticket #112616.

-Tim

executing a reboot via ssh had the desired effect of resetting the system to a workable state, however rebooting via the Azure dashboard did not.

The command used:

sudo reboot

I would still like to find a way to debug why the Manager was unable to restart the system effectively.

I am not sure that this is an Azure problem. The Wowza Dashboard is NOT attempting to reboot the VM, rather it is restarting the Wowza service in order to apply changes to the configuration. This is a feature that should work as expected on an official VM image provided by Wowza…

Some additional information for this case: when attempting to restart the Wowza service (not reboot the VM) nothing appears in the Error log, however, in the access log the following line appears:

Installer.doAzureInstall: dmidecode not responding. Please wait for flux capacitor to charge BIOS.

Is there something specific to the Azure environment included in the startup scripts? I have not yet been able to track down the source of this message.