Fast Snapshot Restore (FSR) is an EBS feature that eliminates the I/O latency penalty when creating a new volume from a snapshot. Without FSR, a restored volume lazy-loads its data — blocks are fetched from S3 on first access, causing slow performance. With FSR enabled, the volume delivers its full provisioned performance immediately upon creation.
Normally when you create an EBS volume from a snapshot, the volume is created instantly but its data blocks are fetched from Amazon S3 in the background as they are accessed for the first time. This lazy-loading behavior causes high latency and low IOPS on first access — which is a serious problem for databases and production workloads that need immediate full performance. Fast Snapshot Restore pre-warms the data so volumes are fully initialized from the moment of creation.
When FSR is enabled on a snapshot for a specific AZ, AWS pre-initializes the snapshot data in a dedicated internal cache
Volumes created from FSR-enabled snapshots deliver full provisioned IOPS and throughput immediately — no warm-up required
FSR must be enabled per snapshot, per Availability Zone separately
There is a limit of 50 FSR-enabled snapshots per region by default (can be increased via AWS Support)
FSR has an additional hourly cost per snapshot per AZ (~$0.75 per snapshot per AZ per hour)
Without FSR — Volume is created instantly but I/O performance is degraded on cold blocks. A workaround is to pre-warm the volume using dd or fio to read all blocks before production use.
Without FSR (workaround) — Use the dd command: sudo dd if=/dev/xvdf of=/dev/null bs=1M to pre-warm all blocks before putting the instance in service.
With FSR — Volume is fully initialized at creation. No pre-warming needed. Immediate full IOPS and throughput. Critical for Auto Scaling scenarios.
Best for — Auto Scaling Groups that need to launch instances quickly with pre-populated data volumes, disaster recovery scenarios, and test environments restored from production snapshots.