Best Practice for Disk Formatting (ReFS vs. NTFS) when Splitting Exchange Server DAG Databases (DB and Logs on same disk) on OCI

Ahmed Essam 211 Reputation points

Hello everyone,

​I am planning a storage optimization task for our Exchange Server environment hosted on Oracle Cloud Infrastructure (OCI) and would appreciate your expert advice on best practices regarding file system selection.

​__Current Environment:__

  • ​__Architecture:__ Exchange Server Database Availability Group (DAG) consisting of 4 nodes (2 in HQ / primary site, 2 in Disaster Recovery / DR site).
  • ​__Storage Layout:__ Currently, all databases and logs are sitting on a single disk/volume per server.
  • ​__Platform:__ Hosted on Oracle Cloud Infrastructure (OCI) compute instances.

​__The Plan:__

To improve performance and distribution, I am adding multiple block volumes to each server to split the databases. Each new disk will host one database and its corresponding log files together (co-located on the same volume).

​__My Question:__

What is the current Microsoft and industry best practice for formatting these new disks for Exchange Server when hosting both DB and logs on the same drive? Should I format them using ReFS or stick with NTFS?

​Are there any specific performance or allocation considerations when combining DB and logs on ReFS within cloud infrastructures like OCI?

​Thank you in advance for your recommendations

0 comments No comments
Sign in to comment

2 answers

  1. Hin-V 15,490 Reputation points • Microsoft External Staff • Moderator

    Hi @Ahmed Essam

    Thank you for sharing your concern. 

    First, I’d like to clarify that this is a user‑to‑user support forum. Moderators participating here do not have access to backend systems, nor can we directly intervene in Microsoft product functionality. Our role is limited to providing technical guidance and sharing best‑practice recommendations based on reported issues, requests, and scenarios.     

    Regarding to your concerns:  

    Best practice for formatting these new disks for Exchange Server 

    Based on my research, ReFS is supported and often recommended for Exchange Server 2016/2019 and later, particularly for volumes hosting database (.edb) files. However, NTFS remains fully supported and continues to be a very safe and widely used choice, especially in scenarios where databases and logs are co-located on the same volume. 

    As your description, where each block volume contains one database along with its corresponding log files, the workload introduces a mixed I/O pattern consisting of sequential log writes and random database I/O. In this type of setup, both NTFS and ReFS are technically viable. That said, NTFS is generally known to provide more predictable and consistent behavior in mixed workloads based on real-world deployments. 

    ReFS does offer advantages such as improved resiliency, block cloning for faster reseeding, and reduced fragmentation. These benefits are typically more noticeable in scenarios where database and log volumes are separated, or where features like AutoReseed or lagged database copies are heavily used. While these advantages can still apply in co-located configurations, they are generally less impactful compared to those optimized designs. 

    You can refer via: Exchange Server storage configuration options | Microsoft Learn 

    Specific considerations for ReFS with DB and Logs on OCI Cloud 

    As a moderator, my support is limited to Microsoft technologies. I do not have direct visibility into third-party cloud platforms such as Oracle Cloud Infrastructure (OCI), including their specific storage performance characteristics. Therefore, I’m unable to provide definitive performance guarantees for ReFS versus NTFS within that environment. 

    That said, in most cloud-based deployments, storage performance is primarily determined by the underlying block volume characteristics, such as provisioned IOPS, throughput, and performance tier, rather than the file system itself. Due to this abstraction layer, the practical performance difference between NTFS and ReFS is usually minimal compared to proper sizing and configuration of the storage. 

    Please note that this summary is based on my own findings and may not fully address your concerns. To help you reach your goal more effectively, I recommend engaging with [GitHub Community Forum] for a deeper technical dive or to connect with individuals who have relevant experience and expertise. Some approaches may behave differently or be restricted depending on your specific environment and configuration. These forums include many experienced developers and Microsoft specialists who can assist with troubleshooting and guidance.     

    I hope this helps. 

    If you have any additional concerns, feel free to comment below. I would be more than happy to assist.

    Note: Please follow the steps in our documentation to enable e-mail notifications if you want to receive the related email notification for this thread. 

    1. Hin-V 15,490 Reputation points • Microsoft External Staff • Moderator

      Hi @Ahmed Essam

      Good day and I hope you're doing well. I just wanted to check in and see if there is any update from you.

      If you have any questions, please feel free to reach out. I’ll be happy to assist further.

      Looking forward to your update.

    Sign in to comment

  2. AI answer

    For Exchange Server DAG databases and logs on OCI block volumes, both NTFS and ReFS are supported. Current Microsoft guidance and the preferred architecture favor ReFS for data volumes, with specific configuration requirements.

    Key points and best practices:

    1. File system choice (ReFS vs. NTFS)
    • Both NTFS and ReFS are supported for Exchange database and log volumes.
      • “File system … Supported: NTFS and ReFS” for Mailbox server roles in both stand-alone and high availability configurations.
    • The Exchange Server preferred architecture specifies using ReFS for all disks that house Exchange data, with integrity disabled:
      • “All disks that house an Exchange data are formatted with ReFS (with the integrity feature disabled) and the DAG is configured such that AutoReseed formats the disks with ReFS.”
    • Therefore, for new DB+log volumes in a DAG, ReFS is the recommended best practice, provided the platform and OS support it correctly.
    1. ReFS configuration for Exchange When using ReFS for Exchange database and log volumes:
    • ReFS is supported for volumes containing Exchange database files, log files, and content index files.
    • Best practice: disable ReFS data integrity features for Exchange database (.edb) files or for the entire volume hosting them:
      • “Data integrity features must be disabled for the Exchange database (.edb) files or the volume that hosts these files.”
    • ReFS allocation unit size:
      • All sizes are supported, but best practice is 64 KB for both .edb and log file volumes.
    1. NTFS configuration (if NTFS is used instead) If NTFS is chosen instead of ReFS:
    • NTFS is fully supported for Exchange database and log volumes.
    • Best practices:
      • NTFS allocation unit size: 64 KB for both .edb and log file volumes.
      • NTFS defragmentation is supported but “not required and not recommended”; on Windows Server 2012, automatic disk optimization/defragmentation should be disabled.
      • NTFS compression is not supported for Exchange database or log files.
    1. DB and logs on the same volume
    • The guidance and best practices for file system and allocation unit size apply equally whether databases and logs are on the same volume or separate volumes.
    • For both NTFS and ReFS, the best practice remains 64 KB allocation unit size for volumes hosting .edb and log files.
    1. Disk and volume layout considerations Independent of NTFS vs. ReFS, for DAG data volumes:
    • Use basic disks rather than dynamic disks as a best practice.
    • Use GPT partitions as a best practice for large volumes.
    • Volume path can be drive letter or mount point; if using mount points, the host volume should be RAID-enabled.
    1. Cloud/OCI-specific note
    • The provided guidance is OS- and Exchange-level and applies equally when the underlying storage is cloud block storage (such as OCI) as long as it is presented as standard Windows disks/volumes.
    • There are no additional Exchange-specific performance or allocation changes required solely because the disks are in a cloud environment; follow the same ReFS/NTFS and allocation unit size best practices.

    Summary:

    • For new DAG volumes hosting both Exchange databases and logs on OCI, the best practice is to format them with ReFS, disable ReFS integrity for database volumes, and use a 64 KB allocation unit size. NTFS with 64 KB allocation unit size remains fully supported if ReFS is not desired or not feasible.

    References:

    AI-generated content may be incorrect. Read our transparency notes for more information.
Sign in to answer

Your answer