How snapshot copy retention interval affects capacity usage in ONTAP?
Applies to
ONTAP 9
Answer
- Snapshot capacity usage fundamentals in ONTAP
- ONTAP Snapshot copies are point in time (PIT) images.
- Snapshot creation does not copy active data blocks.
- Snapshot copies store metadata references to existing blocks.
- When data changes after a Snapshot copy is created:
- Pre change blocks are preserved for Snapshot access.
- New data is written to new blocks by WAFL.
- A data block remains allocated while referenced by one or more Snapshot copies.
- Note: Snapshot capacity usage depends on the amount of changed data since the previous Snapshot copy, not on the total dataset size.
- Capacity comparison: hourly versus daily Snapshot copies
- Example assumptions
- Data change rate: 100 MB per hour
- Total daily change: 2.4 GB
- Each data change overwrites unique blocks
- Hourly Snapshot copies (24 hour retention)
- One Snapshot copy created per hour
- Each Snapshot copy references approximately 100 MB of changed blocks
- 24 Snapshot copies retained
- Approximate Snapshot capacity usage:100 MB × 24 ≈ 2.4 GB
- Daily Snapshot copies (1 day retention)
- One Snapshot copy created per day
- Snapshot copy references all changes from the previous 24 hour
- Approximate Snapshot capacity usage:≈ 2.4 GB
- Result:When block changes do not overlap, hourly and daily Snapshot copies consume approximately the same capacity.
- Example assumptions
- Why Snapshot capacity usage differs in real environments
- Snapshot generation count
- Hourly schedules retain more Snapshot generations.
- More generations extend the lifetime of referenced blocks.
- Extended block retention increases Snapshot space usage.
- Overwrite heavy workloads
- Frequent overwrites cause multiple versions of the same blocks to be retained.
- Shorter Snapshot intervals retain more historical block versions.
- Result:Longer Snapshot intervals typically consume less Snapshot capacity for overwrite heavy workloads.
- Snapshot generation count
- Impact of long retention with frequent Snapshot copies
- Example:
- Hourly Snapshot copies retained for 14 days
- Total Snapshot generations:24 × 14 = 336
- Impact:
- Each Snapshot copy can reference unique or repeatedly changed blocks.
- Blocks cannot be freed until the referencing Snapshot copies expire.
- Snapshot capacity usage increases as:
- Snapshot frequency increases
- Retention duration increases
- Example:
- Effect of increasing the Snapshot interval
- Increasing the Snapshot interval (for example, from hourly to every 2 or 3 hours) can:
- Reduce the total number of Snapshot copies
- Shorten the reference lifetime of old blocks
- Reduce Snapshot capacity usage
- Note: Actual capacity savings vary based on workload type and data change patterns.
- Increasing the Snapshot interval (for example, from hourly to every 2 or 3 hours) can:
