500.0 Error Found in Azure if we access through CName enabled domain name

Question

Hi,

I have a cloudapp.net website for example, http://abc.cloudapp.net

I create a domain called www.testing.com and assign the CName to point http://abc.cloudapp.net.

I am using blobs, App Fabric for session.

When i upgrade or donot access the site for some hours and then I come and access the www.testing.com, I am getting the error as follows:

HTTP 500 - Internal server error.

The interesting part is if I access http://abc.cloudapp.net in a browser and then again reload the www.testing.com, I will get the application loaded.

I also using the setting useBlobMode="true" in appfabric session in the web.config.

I assume that I am missing some accessconfiguration for domain to start the appfabric or blob when it is in Idle mode or Is it because the SQL Azure is Idle?

Can any one please help me on this?

Thanks

S Nagarajan

Answer 1

Do you have a retry policy in place? This is a recommended best practice as requests can and do fail periodically, especially as resources shift around within the datacenter.

That said, I'm not certain what your specific isue is here. I'd recommend you try to drill into the role instances and see if you can capture the exact error message that's causing the http 500 response. This will help isolate if its the app itself, appfabric, or some other dependency. Once we that that info, it should be a bit easier to pinpoint the issue.

Answer 2

Hi,

Thanks for your answer.

For example:

My website name : www.testing.com

My Cloudapp URL : abc.cloudapp.net

My problem is when I access through my website name after an upgradation or certain hours. I am getting an http 500 error.

After that I access the cloudapp URL, my website is loaded and after that i access my website and I am not getting the http error.

I assume now I am clear on the problem.

Thanks

S Nagarajan

Answer 3

You're clear on the symptom, not on the problem. The symptom is the 500 error.

What I'm wondering is if you remote into the VM instance, you can find anything in the logs (either event or IIS) that indicate what the failure is. Maybe you can even see that the instance(s) aren't even getting the initial request, pointing to something within the Azure Fabric or the internet connection that's causing the problem.

A 500 error is fairly generic and could be caused by a million things (I've even seen browsers cache it on rare occasions). So without knowing more about the details behind the 500 error, I'm not sure I can really point you in any direction.