New Azure Batch capabilities
5. september 2016
In the last few weeks, we have added some important new capabilities to the Azure Batch service:
- Batch pool support for Linux VMs and VM scale sets in general availability:
The support for Linux and Windows virtual machines that was made available as preview in April is now generally available. Pool allocation performance and robustness for large numbers of VMs were improved during the preview period.
- Linux MPI:
In December 2015, support for MPI applications was added for Windows VMs. We have now made MPI support available for Linux VMs. For more information, see the blog post.
- Task application packages:
In addition to the ability to specify a list of application packages for a pool, it's now possible to specify an application package list for tasks. For more information, see the MSDN documentation and the Azure article.
- Virtual network support for Windows VMs:
All VMs in a pool can now be part of an Azure virtual network. More information is available in MSDN documentation.
- Job auto-completion:
A property can now be set on a job that enables the job to be automatically moved to the completed state when all tasks associated with the job have finished. Prior to this feature, all tasks had to be monitored and the job explicitly terminated to mark it as completed. For more information, see the documentation for adding a job.
- Persisted job and task output to Azure Storage:
Store job and task output files and log files to the linked Azure storage account by using a new .NET class library. The persisted files can be listed and viewed in the Azure portal. For more information, see the Azure article.
- Azure portal:
We continue to update the Batch capabilities in the Azure portal on a weekly basis, with improved functionality and monitoring capabilities.