Waiting for role to start

Aug 8, 2011 at 9:57 AM

The DotNetNuke Azure Accelerator 6.0 deployment stays forever in the transitioning state with the web-role instance in the waiting for role to start state. Has anyone else encountered this issue?

Aug 8, 2011 at 3:48 PM

Hi @Awwab,

could you check the WADLogsTable for more info?


Aug 9, 2011 at 6:06 AM

Hello David,

Thanks for the reply. I checked the log tables but didn't find anything alarming. In the Infrastructure Log table, the Windows HTTP service was failing when it could not obtain the name of the proxy server with this description: WinHttpGetProxyForUrl(http://dnn0808115.table.core.windows.net) failed ERROR_WINHTTP_AUTODETECTION_FAILED (12180). The Event log table contains errors about the Windows Licensing service with this description and raw XML: The SLUINotify service terminated with the following error: The specified module could not be found.    <Event xmlns='http://schemas.microsoft.com/win/2004/08/events/event'><System><Provider Name='Service Control Manager' Guid='{555908D1-A6D7-4695-8E1E-26931D2012F4}' EventSourceName='Service Control Manager'/><EventID Qualifiers='49152'>7023</EventID><Version>0</Version><Level>2</Level><Task>0</Task><Opcode>0</Opcode><Keywords>0x80000000000000</Keywords><TimeCreated SystemTime='2011-08-09T05:20:12.000Z'/><EventRecordID>3011</EventRecordID><Correlation/><Execution ProcessID='0' ThreadID='0'/><Channel>System</Channel><Computer>RD00155D328272</Computer><Security/></System><EventData><Data Name='param1'>SLUINotify</Data><Data Name='param2'>%%126</Data></EventData></Event>.

Please correct me if I am wrong in thinking that these errors could not be the reason why the web-role never comes out of the waiting for role to start state. Any further pointers would be greatly appreciated.

Aug 10, 2011 at 9:01 PM

Hi Awwab,

which package are you using? The DNN Single and ExtraSmall or the SMB worker role?

Aug 12, 2011 at 2:01 PM

I've Same Problem Im Using SMB Worker Role

Aug 15, 2011 at 4:28 PM

After two successfull deployments, I've built a list about the main reasons of this error. Check the "Documentation" page to see the list and tell me if one of them fix your problem. (http://dnnazureaccelerator.codeplex.com/documentation)

Thanks for feedback

Aug 17, 2011 at 9:05 AM

@jbrinkman found the reason of this bug. The 6.0 Accelerator is not creating the VHD blob container dinamically (see the setting "driveContainer" on the settings configuration file). While this problem is fixed, you can create it manually. 

Aug 17, 2011 at 10:42 AM
Edited Aug 17, 2011 at 10:45 AM

Do we need to create the VHD file manually and then upload it to the blog?  Can I download the VHD file from this site?


Aug 17, 2011 at 11:44 AM
Edited Aug 17, 2011 at 11:50 AM

Hi @bennobright,

It means that you only need to create the "folder" container manually, the VHD will be created at runtime. The default folder container name is "azure-accelerator-drives" as is specified in the service configuration file, so you need to create manually this blob container (create it as a private blob container).

Check this workitem for more info: http://dnnazureaccelerator.codeplex.com/workitem/403

Aug 18, 2011 at 10:44 AM

The package has been fixed and is available for download now.

Sep 26, 2011 at 10:06 AM
Edited Sep 26, 2011 at 1:49 PM

Hi David,

Since Friday I have 2 problems on 2 different "Hosted Service" with the accelerator (v5) :

* On the first project, I had a error 500, so I decided to reboot the web role instance but it didn't fix anything, then I tried to reimage the instance.... But now the status is "waiting for role to start". And in the error table I have "ERROR_BLOB_ALREADY_EXISTS". You talked about this problem in the documentation but not about the solution ?

* For another project I had to install the accelerator on a new hosted service. everything works until the start of the project... Like the previous one, the status is blocked on "waiting for role to start". In the log table I have this error.

I don't know if it's a problem with the accelerator or with Azure because the first project went down for no reason and the second was a new project...

The 2 projects have a VHD, which are mount on the SMB role... So... ? :(

Thanks !

Sep 27, 2011 at 2:56 PM

It seems to be a problem with .. snapshots ?! So the blob container is in read only...

Sep 28, 2011 at 1:20 PM
Edited Sep 28, 2011 at 1:32 PM

Okay.. this morning I accidentally clicked on "delete" instead of "stop" :))))) So I spent my time searching the old package... I found the good one and uploaded it and it works! The webrole has started normally?!

But.... It seems the database was reset to default... Every pages and modules were lost, except the portals list. WHY?! :(


PS: No, I didn't touch it. :)

Sep 28, 2011 at 3:22 PM
Edited Sep 28, 2011 at 3:34 PM

Hi @fthouvein,

Sorry for my late response, I've been travelling these days and I had no time for answer. I have time now so lets try to fix your problem. 

For the sympthoms, seems that Windows Azure replaced your SMB worker role for a new one for maintenance reasons (or you have used the "Upgrade" button on the administration console). The first situation is not common, but can happen (I suffered the same situation on one installation a few months ago). The problem comes when the new worker role want to attach the VHD and the underlying Azure cloud drive API creates a new "read only" snapshot, so finally the SMB worker role uses a read-only drive. The quick solution at this moment is to delete the deployment and deploy again with the same package and service configuration file, so the new deployed worker role will attach the original VHD and not the snapshot. I'm just working in the long term solution for High Availability using leasing and leader election, that I hope that will be available in the next weeks.

The strange thing is the second thing that you mention about the lost pages and modules, except the portals list. The Azure package doesn't execute any SQL statement, only the wizard while creating the new database, so it seems a common DNN problem. I'll try to find some info about that. If you want, contact me via the contact form and perhaps I can help you via Skype or similar in order to see the database contents.



PS: what about the second project? Did you fix the problem?