ConflictError : Windows Azure is currently performing an operation on this deployment that requires exclusive access.

[PraveenSam]

I am trying to add multiple roles(VM's) to the same cloud service and i keep getting the error "ConflictError : Windows Azure is currently performing an operation on this deployment that requires exclusive access." i gave a delay of one minute between each VM creation and it worked. Is it possible to create multiple VMs in the same cloud service parallelly (without giving delays)? And also can i create dedicated cloud service for each VM creation , will there be a limit in the cloud services that we use ?

[bueti]

Hi PraveenSam, the Azure API does currently not support parallel creation of VMs in a single Cloud Service.
There is a limit of number of Cloud Serivces you can create which can be increased via Azure support. You can see you current quota in the Settings panel / Usage on the Azure management portal.

For further questions you might want to head to Stack Overflow (http://stackoverflow.com/questions/tagged/azure) for general questions about Azure.
-Ben

[PraveenSam]

Thanks a lot !!! Thanks for the link too. Sorry , i have one more question. I am currently creating one VM per cloud service , say if i gave 10 request parallely , nearly 7-8 get succeeded , rest 2-3 gets error msg like this :

ConflictError : Windows Azure is currently performing an operation on this hosted service that requires exclusive access.
ConflictError : The operation cannot be performed at this time because a conflicting operation is underway. Please retry later.
ConflictError : The specified DNS name is already taken.

At that instance , i delete that Cloud Service/VM and re-create again. Is there any other workaround ?

[swatisinghi]

Hi PraveenSam, so you are saying that you were not able to create multiple cloud services with one VM each in parallel too?

[PraveenSam]

Hi Swatisinghi,
Yes , if i give request for creation of 10 cloud services parallely with one VM each. Nearly 7-8 get succeeded , rest 2-3 gets error message like this :

ConflictError : Windows Azure is currently performing an operation on this hosted service that requires exclusive access.

ConflictError : The operation cannot be performed at this time because a conflicting operation is underway. Please retry later.

ConflictError : The specified DNS name is already taken.

i call this url -> https://management.core.windows.net//operations/ to get the operation status for every 5-10 seconds to see if the creation of VM/Cloud service has succeeded or not , during which i get some of these errors listed above

[baldwinSPC]

Jumping into this thread as, while I am aware of this issue, I wonder if anyone knows if Microsoft will be resolving this short coming. It seems rather silly that one cannot build out, say, 5 machines at once versus having to do them serially. Do other providers have this same limitation?

[PraveenSam]

Thanks for the info. I have used amazon. It doesn't have this limitation. I can create as much as i want parallelly when i use amazon provider. I guess even Google cloud provider doesn't have this limitation. So you are saying that Microsoft has this limitation ?

[PraveenSam]

It happens for me while i create a seperate cloud service for each VM parallelly.

[baldwinSPC]

You will also see this issue with various other operations, e.g. delete,
endpoint additions, etc. Azure requires most requests to be serialized.
It'd be nice to have an API that allows parallelization of operations
against an instance or cloud service.

On Wed, Jun 25, 2014 at 4:42 AM, PraveenSam [email protected]
wrote:

It happens for me while i create a seperate cloud service for each VM
parallelly.

—
Reply to this email directly or view it on GitHub
#24 (comment)
.

[PraveenSam]

Yes that's right , it happened even during delete too. Thanks @baldwinSPC for confirming.

[swatisinghi]

Yes, this happens with many operations. From what I understood in a discussion with a PM from Azure is that they take lock on resources. They are working on removing this limitation, mean while they suggest to catch the errors and retry!

Google cloud (GCE) and AWS sure doesn't have this limitation.

[PraveenSam]

Thanks @swatisinghi for confirming. Yes i do catch these errors and retry.

[PraveenSam]

The problem for me is that i could create around 1000-2000 VM's parallelly and if i keep catching these errors and retry , it would take a longer time to create all these VM's but i guess there is no other work around.

[baldwinSPC]

@PraveenSam You'll want to grab the operation id and then check to ensure
that ID successfully provisioned before moving onto the next request versus
trying and catching an exception then retrying. I have also noticed that
sometimes a request gets lost, i.e. you submit an add_role request to build
an instance and it disappears in the ether.

On Fri, Jul 4, 2014 at 3:14 AM, PraveenSam [email protected] wrote:

The problem for me is that i could create around 1000-2000 VM's parallelly
and if i keep catching these errors and retry , it would take a longer time
to create all these VM's but i guess there is no other work around.

—
Reply to this email directly or view it on GitHub
#24 (comment)
.

[Rakatashii]

There's no activity going on on my account and I am receiving this message. I updated my cosmos regions, but the operation finished about 30 minutes ago. Not understanding why this is happening, I just want to add a few IP's to the firewall so I can start the application...