Unable to get certain node types allocated

Question

Unable to get certain node types allocated

Erik Heeren 20

Hi,

We're experimenting with different node types on Batch:

Standard_D96as_v5
Standard_D96as_v6
Standard_D96als_v6
Standard_D96ls_v6
Standard_D96s_v6

I can't allocate any of these nodes in east-us or east-us2, instead I get a rather vague

Message: {'additional_properties': {}, 'lang': 'en-US', 'value': 'Server encountered an internal error. Please try again after some time.\nRequestId:6f275e2d-9b6f-4f99-8d41-90d120762937\nTime:2025-07-21T11:29:38.4397555Z'}

I'm sure I don't need to specify that trying again later doesn't help.

What's actually going wrong here?

Himanshu Shekhar 1,170 Reputation points Microsoft External Staff Moderator

2025-08-18T18:43:54.7233333+00:00
Hi Erik Heeren

Welcome to Microsoft Q&A Platform. Thank you for reaching out & hope you are doing well.

Currently, you may not allocate Standard_D96as_v5, D96as_v6, D96als_v6, D96ls_v6, or D96s_v6 nodes in east-us or east-us2 due to limited regional capacity for these VM SKUs.

These 96-core SKUs are in high demand and often face capacity constraints, especially in popular regions like east-us and east-us2. Even with sufficient quota, Azure may not have the available physical resources at this time.

The error indicates that Azure does not have enough available hardware to support the requested VM type in your region, or that the SKU is currently restricted or unavailable for Batch in that region.

Possible try out:

Try different VM sizes/SKUs, especially smaller sizes or other series.

Try other regions where capacity may exist.

If you must use these SKUs in east-us/east-us2, consider requesting a capacity reservation.

Please refer Microsoft documentation for more details:

Troubleshoot allocation failures when you create or resize VMs in Azure: Troubleshooting Azure VM allocation failures - Virtual Machines | Microsoft Learn

Azure batch pool resizing failure : Azure batch pool resizing failure - Azure | Microsoft Learn

Azure batch Reliability in Azure Batch | Microsoft Learn

Also please refer the details as well suggested by our MVP.

Kindly let us know if the suggested details helps or you need further assistance on this issue.

Regards

Himanshu

Answer accepted by question author

0 additional answers

Your answer

Himanshu Shekhar 1,170 Reputation points Microsoft External Staff Moderator

2025-08-18T18:43:54.7233333+00:00

Hi Erik Heeren

Welcome to Microsoft Q&A Platform. Thank you for reaching out & hope you are doing well.

Currently, you may not allocate Standard_D96as_v5, D96as_v6, D96als_v6, D96ls_v6, or D96s_v6 nodes in east-us or east-us2 due to limited regional capacity for these VM SKUs.

These 96-core SKUs are in high demand and often face capacity constraints, especially in popular regions like east-us and east-us2. Even with sufficient quota, Azure may not have the available physical resources at this time.

The error indicates that Azure does not have enough available hardware to support the requested VM type in your region, or that the SKU is currently restricted or unavailable for Batch in that region.

Possible try out:

Try different VM sizes/SKUs, especially smaller sizes or other series.

Try other regions where capacity may exist.

If you must use these SKUs in east-us/east-us2, consider requesting a capacity reservation.

Please refer Microsoft documentation for more details:

Troubleshoot allocation failures when you create or resize VMs in Azure: Troubleshooting Azure VM allocation failures - Virtual Machines | Microsoft Learn

Azure batch pool resizing failure : Azure batch pool resizing failure - Azure | Microsoft Learn

Azure batch Reliability in Azure Batch | Microsoft Learn

Also please refer the details as well suggested by our MVP.

Kindly let us know if the suggested details helps or you need further assistance on this issue.

Regards

Himanshu

Answer 1

Hi ,

Thanks for reaching out to Microsoft Q&A.

Nothing is wrong with your Batch pool setup. The SKUs you are requesting are simply not in capacity in East US / East US 2 today. That “internal error” is the symptom of azure having no supply of those VM types in those regions. The real fix is to either (a) move to a region where the SKUs exist, or (b) request capacity confirmation from Microsoft.

Behind the scenes:

The node families you listed are the newer Ampere/Zen4 SKUs (D96as_v5, D96as_v6, D96als_v6, D96ls_v6, D96s_v6). These are very large 96-core machines. They are only deployed in limited clusters within each region.
Batch pools do not “magically” pull a VM SKU from anywhere: they are constrained to the underlying scale unit capacity in that region + zone. If the SKU is not stocked in East US / East US 2, or if the cluster has no free quota, the allocation call fails.
The message you see (Server encountered an internal error) is unfortunately a generic “capacity error” that comes back when Batch asks the Fabric Controller for VMs and it responds “none available”.
- When it is truly a quota issue (you have not been approved for that SKU in that region), you normally see a clearer error (AllocationFailed, CoreQuotaExceeded).
- When it is zero capacity in the cluster (the SKU is either not deployed yet or fully oversubscribed), it surfaces as this vague “internal error”.
That is why “trying again later” does not help: there is simply no inventory of those SKUs in east-us/east-us2 right now.

How to verify and work around:

Run az vm list-skus --location eastus --size Standard_D96 --all to check if those SKUs are actually offered in that region. You will likely see them marked as NotAvailableForSubscription or Unavailable.

Try other regions where Microsoft has stocked the v6 SKUs (for example: westus3, southcentralus, or canadacentral tend to get the new silicon first).

If you must use east-us, open a support request with azure to check if and when those SKUs will be made available there. Azure sometimes staggers rollout of v6 SKUs region by region.
If your workload is flexible, consider smaller sizes in the same family (ex: D48as_v6, D32as_v6) these are often more readily available.

Please 'Upvote'(Thumbs-up) and 'Accept' as answer if the reply was helpful. This will be benefitting other community members who face the same issue.

Erik Heeren 20 Reputation points

2025-08-19T10:02:44.7366667+00:00

Hi Himanshu Shekhar* and @Vinodh247 ,

Thanks for getting back to me and clarifying! It seems like for the most part the nodes not being present in the region is indeed the cause.

However let's take a look at e.g. Standard_D96alds_v6: (capabilities stripped for readability)

{ "family": "standardDaldv6Family", "locationInfo": [ { "location": "eastus", "zoneDetails": [ { "capabilities": [ { "name": "UltraSSDAvailable", "value": "True" } ], "name": [ "3", "1", "2" ] } ], "zones": [ "3", "1", "2" ] } ], "locations": [ "eastus" ], "name": "Standard_D96alds_v6", "resourceType": "virtualMachines", "restrictions": [], "size": "D96alds_v6", "tier": "Standard" },

No restrictions here, available in the region I'm trying in, and yet it results in the same internal server error. Does that mean there are nodes of this type available, but simply reserved / in use by other customers?

At this point I'm just trying to increase my understanding of Azure, az vm list-skus --location eastus --size Standard_D96 --all was a big help!

*I would tag you too, but there are at least 30 accounts with the same name and no way to tell which one is you.
Erik Heeren 20 Reputation points

2025-08-20T09:40:23.6833333+00:00

I had some time so I went through the list of Standard_D96 SKUs which should be available in the us-east region (no restrictions, appear in the list with the above command).

Of those, my pool create command failed for:

Standard_D96ads_v6

Standard_D96alds_v6

Standard_D96ds_v6

Standard_D96lds_v6

Standard_D96pds_v6

Standard_D96plds_v6

Standard_D96pls_v6

Standard_D96ps_v6

This was the error message:

Message: {'additional_properties': {}, 'lang': 'en-US', 'value': 'The value provided for one of the properties in the request body is invalid.\nRequestId:4cfd05a1-dff6-46f2-9a7a-3bc55857dbb7\nTime:2025-08-20T09:33:19.0031661Z'} PropertyName: vmSize PropertyValue: Standard_D96ads_v6

When trying to create my pool through the portal instead of the API, I was unable to find any of these vm sizes there.

Which leads me to the question: what does it mean that they're present in the output of az vm list-skus --location eastus --size Standard_D96 --all but that they're not present when trying to select a vm size in the portal?

Share via

Unable to get certain node types allocated

0 additional answers

Your answer