Skip to content

Add space.nearly_full_threshold_percent to ONTAP NAS backend config defaults #1171

Description

@mcnairda

Is your feature request related to a problem? Please describe.

When is-space-enforcement-logical = true is set on Trident-provisioned ONTAP NAS volumes (required to prevent VM pause events on OpenShift Virtualization under sustained write load), ONTAP activates internal per-volume logical space tracking. ONTAP's space monitor (monitor.vol.nearFull.inc.sav) watches this tracking and fires at the default 95% nearFull threshold for every Trident PVC continuously — because those volumes are designed to fill under logical enforcement. This generates thousands of alerts across large environments (40+ clusters, thousands of PVCs) with no native mechanism to set the threshold at creation time.

Describe the solution you'd like

space.nearly_full_threshold_percent is a valid ONTAP REST API parameter on volume creation (POST /api/storage/volumes) and modification (PATCH /api/storage/volumes/{uuid}). Trident does not currently expose it in its backend config defaults section.

Adding this as a configurable parameter in TridentBackendConfig would allow operators to set the nearFull threshold at PVC creation time — for example, spaceNearlyFullThresholdPercent: 99 — so every new PVC inherits the correct threshold with zero ongoing effort across any number of clusters. This eliminates the need for post-provisioning scripts, EMS route suppression, or recurring scheduled jobs to manage this at scale.

Describe alternatives you've considered

Post-provisioning PATCH script on a schedule — calls PATCH /api/storage/volumes/{uuid} with space.nearly_full_threshold_percent=99 after Trident creates the volume. Operational overhead is not acceptable at 40+ clusters with continuous PVC provisioning.
ONTAP EMS webhook (ONTAP 9.9+) — triggers an external HTTP endpoint on vol.create events, which then calls the ONTAP REST API to apply the threshold. Event-driven and requires no polling, but introduces an external service dependency that must be maintained and secured across all clusters.
EMS route severity suppression — downgrades monitor.vol.nearFull.inc.sav to notice cluster-wide, silencing the alerts entirely. Pragmatic at scale but eliminates per-volume capacity visibility completely rather than tuning it appropriately.
vserver modify — confirmed that space-nearly-full-threshold-percent is not available as an SVM-level default. Volume-only parameter.
Additional context

This request stems directly from the recommended ONTAP configuration for OpenShift Virtualization environments using Trident NAS backends. Setting is-space-enforcement-logical = true is the confirmed fix for VM pause events caused by WAFL COW overwrite transients on per-PVC FlexVols. However, enabling logical enforcement activates ONTAP's .inc.sav logical space monitor, which fires at the default 95% threshold against every Trident PVC that fills under the new enforcement model. At enterprise scale (40+ ONTAP clusters), managing this threshold post-creation is not operationally viable without native Trident support.

Reference: ONTAP REST API — POST /api/storage/volumes, space.nearly_full_threshold_percent parameter:
[https://docs.netapp.com/us-en/ontap-restapi/ontap/post-storage-volumes.html]

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions