Controlling maxfiles
Suggest changes
PDFs
The volume -files option is the public inode ceiling. Raise it to a planned value, or to the current files-maximum-possible, before the workload runs out of inodes. Limits, NFS file-ID behavior, and exhaustion symptoms are described in Maxfiles and ONTAP inode information.
Set a specific ceiling
Use volume modify -files when capacity planning calls for a specific value:
volume modify -vserver <svm> -volume <volume> -files <inode-count>
ONTAP validates the requested value against volume size, the maximum possible value, and inode-file capacity already allocated. A configured ceiling cannot be reduced below the inode-file capacity already reached.
For a FlexGroup, specify the FlexGroup name and desired FlexGroup-wide total. Do not set files independently on a constituent. ONTAP distributes the requested total evenly among the current constituents, and each constituent enforces its portion:
volume modify -vserver <svm> -volume <flexgroup> -files <flexgroup-total-inodes>
After changing the value, inspect constituent-level files and files-used because usage does not necessarily remain as evenly distributed as the configured limits:
volume show -vserver <svm> -volume-style-extended flexgroup-constituent -fields files,files-used
Some ONTAP commands accept a relative increase:
volume modify -vserver <svm> -volume <volume> -files +<additional-inodes>
Confirm the result with volume show.
Setting files to the maximum up front
Raising files to the current files-maximum-possible does not preallocate the public inode file and does not reserve the corresponding inode-record capacity. Immediate space risk is therefore low. For high-file-count NAS volumes, set files to the maximum the volume can hold whenever volume size and protocol constraints allow it, rather than waiting for an out-of-inodes event.
The remaining implications are about what happens if the namespace actually grows toward that ceiling, and about hitting a limit you cannot raise further:
-
Capacity is consumed only as inodes are allocated. Each ONTAP 9 public inode uses 288 bytes in the inode file, plus inode-file structure and other metadata. At the FlexVol absolute cap, raw inode records alone are about 2,040,109,451 × 288 bytes (approximately 547 GiB). That used space also occupies the underlying storage, so monitor total free space, not only volume used percentage. See Capacity impact.
-
After the inode file has grown, you cannot lower
filesbelowinodefile-public-capacity. Setting the ceiling high does not lock unused capacity; allocated records do. -
On a FlexVol, 2,040,109,451 is a hard stop. You cannot add more public inodes to that volume. If the workload might exceed one volume's absolute cap, use a FlexGroup so you can add constituents. Setting a FlexVol or constituent to the maximum its current size allows is still appropriate; that value is often far below 2,040,109,451.
-
A FlexGroup-wide total can exceed two billion inodes (up to 400 billion in unified ONTAP and up to 1 trillion in AFX running ONTAP 9.19.1 and later). For NFS, enable 64-bit file identifiers before the visible file-ID count approaches two billion. See NFS 64-bit file identifiers and FlexGroup file counts.
-
Continue to monitor
files-usedagainstfiles. A maximum setting removes the default ceiling as a surprise, but the volume can still run out of inodes at the supported cap.
To apply the current maximum, use the advanced files-set-maximum action beginning with ONTAP 9.9.1:
set -privilege advanced volume modify -vserver <svm> -volume <volume> -files-set-maximum true
This is a one-shot action, not a persistent allocation mode. Only true is valid. If the volume later grows, run it again if the new maximum should be applied. Use -files when you want a deliberate value below the maximum. If a FlexVol volume (or FlexGroup constituent) is approximately 7.8 TB or larger, the maximum you can set is the absolute cap of 2,040,109,451, and that value does not increase further with volume size.
Next: Monitor maxfiles, EMS events, and ONTAP enhancements → |