Loading...

Accelerated Networking with H-series VMs on Azure for older OS distributions

Accelerated Networking with H-series VMs on Azure for older OS distributions

The accelerated networking update for the HPC SKUs on Azure has caused problems for older OS distributions or any MPI versions that do not use the latest UCX. This is due to inconsistent naming for the IB devices. My recent patch to rdma-core can be used to provide consistent naming with udev rules. This following script can be used when building an image:

 

yum install -y cmake libnl3-devel
git clone https://github.com/linux-rdma/rdma-core.git
cd rdma-core
bash build.sh
cp build/bin/rdma_rename /usr/lib/udev/
cat <<EOF >/etc/udev/rules.d/60-ib.rules
# Accelnet board
ACTION=="add", ATTR{board_id}=="MSF0010110035", SUBSYSTEM=="infiniband", PROGRAM="rdma_rename %k NAME_FIXED mlx5_an0"
# HBv2 board
ACTION=="add", ATTR{board_id}=="MT_0000000223", SUBSYSTEM=="infiniband", PROGRAM="rdma_rename %k NAME_FIXED mlx5_ib0"
# HC board
ACTION=="add", ATTR{board_id}=="MT_0000000010", SUBSYSTEM=="infiniband", PROGRAM="rdma_rename %k NAME_FIXED mlx5_ib0"
EOF

This will name the accelerated networking mlx5_an0 and the infiniband to mlx5_ib0. Now, you can use the older MPI/UCX versions by setting:

 

export UCX_NET_DEVICES=mlx5_ib0:1

The script includes rules that will work for HB, HC, HBv2 and NDv2.

Published on:

Learn more
Azure Global articles
Azure Global articles

Azure Global articles

Share post:

Related posts

Azure Developer CLI (azd) Dec 2025 – Extensions Enhancements, Foundry Rebranding, and Azure Pipelines Improvements

This post announces the December release of the Azure Developer CLI (`azd`). The post Azure Developer CLI (azd) Dec 2025 – Extensions En...

2 days ago

Unlock the power of distributed graph databases with JanusGraph and Azure Apache Cassandra

Connecting the Dots: How Graph Databases Drive Innovation In today’s data-rich world, organizations face challenges that go beyond simple tabl...

4 days ago

Azure Boards integration with GitHub Copilot

A few months ago we introduced the Azure Boards integration with GitHub Copilot in private preview. The goal was simple: allow teams to take a...

5 days ago

Microsoft Dataverse – Monitor batch workloads with Azure Monitor Application Insights

We are announcing the ability to monitor batch workload telemetry in Azure Monitor Application Insights for finance and operations apps in Mic...

6 days ago

Copilot Studio: Connect An Azure SQL Database As Knowledge

Copilot Studio can connect to an Azure SQL database and use its structured data as ... The post Copilot Studio: Connect An Azure SQL Database ...

6 days ago

Retirement of Global Personal Access Tokens in Azure DevOps

In the new year, we’ll be retiring the Global Personal Access Token (PAT) type in Azure DevOps. Global PATs allow users to authenticate across...

9 days ago

Azure Cosmos DB vNext Emulator: Query and Observability Enhancements

The Azure Cosmos DB Linux-based vNext emulator (preview) is a local version of the Azure Cosmos DB service that runs as a Docker container on ...

10 days ago

Azure Cosmos DB : Becoming a Search-Native Database

For years, “Database” and “Search systems” (think Elastic Search) lived in separate worlds. While both Databases and Search Systems oper...

10 days ago
Stay up to date with latest Microsoft Dynamics 365 and Power Platform news!
* Yes, I agree to the privacy policy