New Production Release: Appistry CloudIQ Platform 4.2.0 Community Edition

I have a problem to report with 4.2.0. When I did a upgrade though a fabric_ctl deploy my current applications stop working. When I did a uninstall/install directly on the worker everything works.

I have both environment up and running if more information is desired. I don't really want to do the uninstall/install on all of my workers. I am using 32bit Windows server 2003.

Thanks for the input. In order to narrow down the exact issue, it would be helpful if you could give us a little more detail about the applications that stopped working. Did they fail to load? Are they .NET applications? Did the fabric itself stop working or just the applications? Are you able to get a log_monitor capture on the fabric that is having the issue? Could you issue a stop and start of the application while capturing the log_monitor traffic?

Thanks again.

Alright
Did they fail to load? - By load do you mean deploy the application? The applications fail to run but I can update/deploy the application without error in the UI, I get and messagein the log_monitor

Nov 04 10:29:53 10.3.11.42 fabric_task_service.hello_world_dotnet_app[I:4042] Failed to start the .NET CLR: Error creating .NET Task Handler Factory: Class not registered
Nov 04 10:29:53 10.3.11.42 fabric_task_service.hello_world_dotnet_app[E:4001] HelloWorld task failed to load.  Class: 'Appistry.Samples.HelloWorld.HelloWorld' Method: 'Greet' File: 'hello_world_dotnet_app.dll' XML: 'hello_world_task.xml'  Reason: Error creating .NET Task Handler Factory: The system cannot find the file specified.
Nov 04 10:29:53 10.3.11.42 fabric_task_service.hello_world_dotnet_app[I:1000] Service started on port 4732
Nov 04 10:29:53 10.3.11.42 Seneschal[I:7519] service 'fabric_task_service.hello_world_dotnet_app' running

Appistry.FabricAPI.TransactionException: TaskID(hello_world_dotnet_app.HelloWorld) not found
   at Appistry.FabricAPI.ExceptionHelper.HandleResultCode(ErrorCode rc, FabricRequest request)
   at Appistry.FabricAPI.Fabric.ExecuteRequest(FabricRequest request, SubmitFunction function)
   at Appistry.FabricAPI.Fabric.Execute(FabricRequest fabricRequest)
   at Appistry.Samples.HelloWorld.HelloWorldClient.Main(String[] args)

Nov 04 10:22:16 10.3.11.49 fabric_task_service.hello_world_dotnet_app[I:4042] Failed to start the .NET CLR: Error creating .NET Task Handler Factory: Class not registered
Nov 04 10:22:16 10.3.11.49 fabric_task_service.hello_world_dotnet_app[E:4001] HelloWorld task failed to load.  Class: 'Appistry.Samples.HelloWorld.HelloWorld' Method: 'Greet' File: 'hello_world_dotnet_app.dll' XML: 'hello_world_task.xml'  Reason: Error creating .NET Task Handler Factory: The system cannot find the file specified.

Is the deploy not registering some Appistry dlls?
I'm happy to answer any more questions, but I'm reinstalling appistry by hand on my workers Friday. I need a working cloud

Got ya.

I will work with our development team to find the options that will be the least hassle. I will post here again tomorrow.

Thanks,
Mark

I was able to recreate the issue on our end. The easiest way to resolve the issue is to install the new assembly files with our fabric_gac.exe utility.

To do so, follow this procedure:

• Stop your workers via fabric_ctl or the CloudIQ Console.
• Change directory to the FABRIC_HOME directory. By default, this is the C:\Program Files\Appistry\CloudIQ\system directory.
• Install the new assembly and its policies by typing "fabric_gac.exe -i Appistry.NET.dll"
• Start your workers

I did not have to reboot them for the applications to load, but I wouldn't be surprised if that is necessary.

Please do this and tell us if that resolves the issue.

Thanks,
Mark

Do I have run the fabric_gac.exe on each worker?

Yes, you would have to run it on each worker. Do you have remote desktop enabled on your workers? Maybe that will make it a little easier.

Did you have a chance to install the new Appistry.NET.dll on each worker? If so, did that resolve the issue?

Thanks,
Mark

Yes, I followed the instructions on each worker and everything works now.