What is the Directory Indexing Scanner and what does directory indexing accomplish?
Applies to
- Data ONTAP 9.2 and higher
- Storage System Performance
Answer
As datasets have gotten larger and larger, performance can become a concern on those larger datasets. ONTAP has scaled to meet these needs as it has evolved, and one of those features is directory indexing.
- Starting with ONTAP 9.2, a feature was introduced which will index directory sizes 2MB or larger .
- Directory lookup will only read required blocks in the directory instead of entire directory.
- Loading 100 directories of 10 MB each, prior to ONTAP 9.2, would require 1.25 GB of RAM.
- Indexing the same 100 directories of 10 MB each in 9.2+ reduces this to around 200 MB of RAM.
This will be noticed in a few ways:
- ONTAP upgrade. You will notice in "wafl scan status" a "directory index creation" scan.
- Reboot may trigger a "directory index creation" scan as well.
- A frontend call (lookup/create generally) will cause ONTAP to notice that the directory is greater than 2 MB and index it.
- You may see in waffinity stats/wafl suspend data the WAFL messages related to indexing.
Additional Information