hesela.devONTAP consistency point
In ONTAP, a consistency point writes buffered filesystem changes to storage and, after those writes complete, updates the active filesystem pointers.
The normal write path processes buffered data through WAFL and RAID before changing the on-storage pointers. A consistency point is this filesystem commit step, not the arrival of each individual host write. [1]
Acknowledgment: With both HA partners operating, logging on the local node and its partner can permit acknowledgment before the later consistency point. Recovery can replay protected outstanding writes into memory and commit them through a subsequent consistency point. [1] [2]
Relations
- related
- Non-volatile random-access memory, All-flash array, Data durability, Crash consistency, Write-ahead log
Limitations
- Filesystem consistency does not by itself establish application transaction consistency, and loss of both powered nodes interrupts access. [2]
Citations
- NetApp: What is Consistency Point, and why does NetApp use it? Applies to ONTAP 9 and Data ONTAP 7/8.
- NetApp ONTAP 9: unexpected power outage, protected-write replay and application consistency.
Sources & caveats
- ONTAP-specific filesystem terminology, not a generic name for a distributed snapshot or an application transaction boundary. NetApp's operational documentation is the primary evidence.
- No fixed timer, buffer capacity or throughput is inferred across platforms and releases.