What is the expected data read and write behavior in the event of a hardware issue when using the Locally Redundant Storage option for ADLS?

When using Locally Redundant Storage (LRS) for Azure Data Lake Storage (ADLS), the expected behavior during a hardware issue such as a disk, server, or rack failure is as follows:

1. Data Availability: LRS replicates your data within a single physical datacenter. While it protects against drive, server, and rack failures, if a disaster occurs that affects the entire datacenter (like fire or flooding), all replicas may be lost or unrecoverable.
2. Write Operations: A write request to a storage account using LRS is synchronous, meaning the write operation is considered successful only after the data is written to all three replicas. If a hardware issue occurs, the write operations may be temporarily unavailable until the issue is resolved.
3. Outage Transparency: The outage is not transparent to applications. If the hardware issue affects the availability of the storage account, applications may experience downtime or errors until the issue is resolved.
4. Data Resynchronization: Once the hardware issue is resolved, the data does not automatically resync. The replicas are already synchronized as part of the LRS mechanism, but if there was a temporary unavailability, the system would need to recover from that state, which may involve manual intervention depending on the nature of the failure.

In summary, while LRS provides a level of redundancy, it does not protect against total datacenter failures, and applications may experience downtime during hardware issues.

References:

• Azure Storage redundancy
• Azure Files data redundancy
• Disaster recovery and failover for Azure Files
• Redundancy options for managed disks
• Plan for an Azure File Sync deployment

Hi @Michael Evans

When ADLS uses Locally Redundant Storage (LRS), all data is synchronously replicated three times within a single datacenter across separate fault and upgrade domains.

During a disk, server, or rack hardware failure:

• The platform automatically serves your data from healthy replicas.
• Small outages or retryable errors may occur—failover is not guaranteed to be seamless or transparent.
• Writes require all replicas to commit, so write availability may be affected during hardware repair.

Once hardware recovers, Azure automatically repairs and resynchronizes replicas without application involvement. No customer action is required.

However, LRS does not protect against full datacenter outages. If the entire facility is impacted, all three replicas may become unavailable. For stronger durability or application‑transparent failover, Microsoft recommends ZRS, GZRS, or RA‑GZRS.

References:

• Locally Redundant Storage (LRS)

Hope this helps! Feel free to ask if you have more questions!