I am getting a System.AccessViolationException thrown during the execution of an Azure Web Role (run on the Azure emulator, this has not been uploaded to Azure yet) when a call is made to an overridden method of an object when a local int variable is passed as one of the method parameters. The exception message is "Attempted to read or write protected memory. This is often an indication that other memory is corrupt".
The code where the exception is thrown is part of a local library that has been used for several years on live systems (not Azure) with no issues. The part that errors is as follows:
foreach (XmlDataComponent item in this.items)
{
int index = 0;
XmlNode node = item.ToXml(dataSet, xmlDocument, this, index); // Exception thrown when this call is made
...
}
The XmlDataComponent is a base class, when the code runs item is one of its derived classes. The ToXml() method is overridden in the derived classes. The exception is thrown as soon as the call is made to ToXml().
The problem is the index parameter. If I swap this to use an explicit value instead of the local variable, e.g.
item.ToXml(dataSet, xmlDocument, this, 0)
there are no errors.
Similarly, if I cast the item to its actual type e.g.
((XmlDataItem)item).ToXml(dataSet, xmlDocument, this, index))
and mark the ToXml() method in the XmlDataItem class as new instead of override there are no errors.
I have also tried calling the library from a console application rather than a web role with exactly the same data (i.e. everything the same other than running under a web role). Again, this caused no problems.
It appears that when run under the Azure emulator, accessing a local variable as a parameter to an overridden method is an issue!!!
I'm hoping this is only an issue when run under the emulator, however we still need a fix otherwise dev is more difficult.
Any suggestions or advise would be much appreciated.
Related
When I run (func run) the azure function project locally (VSCode) after I add azure function proxy I'm getting 'Worker was unable to load function" console error. However function locally run perfectly after that console error.
What could be the reason for that console error
Error
Worker was unable to load function get-user: 'TypeError [ERR_INVALID_ARG_VALUE]: The argument 'id' must be a non-empty string. Received '''
The problem seems to be that the GUID of the function is generated but no real instance is generated. Therefore, trying to access the function GUID through proxy will report an error because it does not exist at all. I don't think you need to worry about this. This should be a bug, but the good news is that it will not have any effect on your function, because you will never use this GUID any more. A new instance and corresponding GUID will be generated when you trigger the function, and the previous GUID will be discarded.
I have an azure function that runs off of a queue trigger. The repository has method to grab the connection string from the ConnectionStrings collection.
return System.Configuration.ConfigurationManager.ConnectionStrings["MyDataBase"].ToString();
This works great for the most part but I see intermittently that this returns a null exception error.
Is there a way I can make this more robust?
Do azure functions sometimes fail to get the settings?
Should I store the setting in a different section?
I also want to say that this runs thousands of times a day but I see this popup about a 100 times.
Runtime version: 1.0.12299.0
Are you reading the configuration for every function call? You should consider reading it once (e.g. using a Lazy<string> and static) and reusing it for all function invocations.
Maybe there is a concurrency issue when multiple threads access the code. Putting a lock around the code could help as well. ConfigurationManager.ConnectionStrings should be tread-safe, but maybe it isn't in the V1 runtime.
A similar problem was posted here, but this concerned app settings and not connection strings. I don't think using CloudConfigurationManager should be the correct solution.
You can also try putting the connection string into the app settings, unless you are using Entity Framework.
Connection strings should only be used with a function app if you are using entity framework. For other scenarios use App Settings. Click to learn more.
(via Azure Portal)
Not sure if this applies to the V1 runtime as well.
The solution was to add a private static string for the connection string. Then only read from the configuration if it fails. I then added a retry that paused half a second. This basically removed this from happening.
private static string connectionString = String.Empty;
private string getConnectionString(int retryCount)
{
if (String.IsNullOrEmpty(connectionString))
{
if (System.Configuration.ConfigurationManager.ConnectionStrings["MyEntity"] != null)
{
connectionString = System.Configuration.ConfigurationManager.ConnectionStrings["MyEntity"].ToString();
}
else
{
if (retryCount > 2)
{
throw new Exception("Failed to Get Connection String From Application Settings");
}
retryCount++;
getConnectionString(retryCount);
}
}
return connectionString;
}
I don't know if this perfect but it works. I went from seeing this exception 30 times a day to none.
I am having problems getting the Microsoft.Azure.Documents library to initialize the client in an azure worker role. I'm using Nuget Package 0.9.1-preview.
I have mimicked what was done in the example for azure document
When running locally through the emulator I can connect fine with the documentdb and it runs as expected. When running in the worker role, I am getting a series of NullReferenceException and then ArgumentNullException.
The bottom System.NullReferenceException that is highlighted above has this call stack
so the nullReferenceExceptions start in this call at the new DocumentClient.
var endpoint = "myendpoint";
var authKey = "myauthkey";
var enpointUri = new Uri(endpoint);
DocumentClient client = new DocumentClient(endpointUri, authKey);
Nothing changes between running it locally vs on the worker role other then the environment (obviously).
Has anyone gotten DocumentDb to work on a worker role or does anyone have an idea why it would be throwing null reference exceptions? The parameters getting passed into the DocumentClient() are filled.
UPDATE:
I tried to rewrite it being more generic which helped at least let the worker role run and let me attached a debugger. It is throwing the error on the new DocumentClient. Seems like some security passing is null. Both the required parameters on initialization are not null. Is there a security setting I need to change for my worker role to be able to connect to my documentdb? (still works locally fine)
UPDATE 2:
I can get the instance to run in release mode, but not debug mode. So it must be something to do with some security setting or storage setting that is misconfigured I guess?
It seems I'm getting System.Security.SecurityExceptions - only when using The DocumentDb - queues do not give me that error. All Call Stacks for that error seem to be with System.Diagnostics.EventLog. The very first Exception I see in the Intellitrace Summary is System.Threading.WaitHandleCannotBeOpenedException.
More Info
Intellitrace summary exception data:
top is the earliest and bottom is the latest (so System.Security.SecurityException happens first then the NullReference)
The solution for me to get rid of the security exception and null reference exception was to disable intellitrace. Once I did that, I was able to deploy and attach debugger and see everything working.
Not sure what is between the null in intellitrace and the DocumentClient, but hopefully it's just in relation to the nuget and it will be fixed in the next iteration.
unable to repro.
I created a new Worker Role. Single instance. Added authkey & endoint config to cscfg.
Created private static DocumentClient at WorkerRole class level
Init DocumentClient in OnStart
Dispose DocumentClient in OnStop
In RunAsync inside loop,
execute a query Works as expected.
Test in emulator works.
Deployed as Release to Production slot. works.
Deployed as Debug to Staging with Remote Debug. works.
Attached VS to CloudService, breakpoint hit inside loop.
Working solution : http://ryancrawcour.blob.core.windows.net/samples/AzureCloudService1.zip
While creating an Incident. I am facing this issue. I have created some incidents and it working fine but sometime it shows this exception "Incorrect attribute value type Microsoft.Xrm.Sdk.OptionSetValue".
How to trace which OptionSetValue throws an error.
I'm going to assume that you're creating these via a C# sdk call. I believe turning on server side tracing would tell you. The other option is if you can debug it locally, catch the exception and manually attempt to update each optionset attribute, one at a time until you find the culprit.
I've configured the Azure Diagnostics so that the logs get uploaded to a storage table. I'm using Trace.TraceXxx from my code and all works well.
Now I'm trying to add tracing from the Role OnStart() and OnStop() methods. I know that the tracing works as I see the lines in the Debug window when running in the emulator. But from the cloud deployment, it seems that these trace lines never get uploaded to the table. My guess is that it is somewhat related to TraceSources, as the only trace lines I've in the table come from the w3wp.exe source... Any hint ?
Thanks
Like you said you can add the trace listener using the WaIISHost.exe.config, but besides that you can also add the trace listener in code (you'll need a reference to Microsoft.WindowsAzure.Diagnostics.dll):
public class WebRole : RoleEntryPoint
{
public override void Run()
{
var listener = new Microsoft.WindowsAzure.Diagnostics.DiagnosticMonitorTraceListener();
Trace.Listeners.Add(listener);
...
}
}
Another way of setting up diagnostics is through the configuration file. If you created a VS solution recently, it will automatically create the diagnostics plug-in and configuration for the trace listener. With the config file (diagnostics.wadcfg) there is no code that needs to be written for the different data sources. Here is a link where you can get started and a sample file:
http://msdn.microsoft.com/en-us/library/gg604918.aspx
You cannot include custom performance counters right now and you need to make sure that you don’t try to allocate more than 4GB of buffer to anything (you can leave at 0), or it tends to fail.
Note, the time interval format (e.g PT1M). That is a serialization format, so PTXM is X minutes, while PTXS is X in seconds. You need to mark this as content and copy always in Visual Studio (place at root of project) so it gets packaged.
And here is a link to the three ways to setup diagnostics
http://msdn.microsoft.com/en-us/library/windowsazure/hh411541.aspx
Ranjith
http://www.opstera.com