The application policy created for taking backup via Rubrik showed no further increase in hit count despite active traffic. The customer also experienced low throughput, slowness, and lag during backup transfers. Clearing the application's sessions temporarily restored expected performance.

Investigation indicated that the behavior was associated with long-lived NFS sessions and WAN link utilization, rather than a confirmed application policy hit-count defect.

The application policy hit count remained unchanged while traffic was actively flowing.

Backup transfers experienced low throughput, slowness, and lag.

Performance improved temporarily after clearing the application's sessions.

Traffic was distributed across the ISP1 and ISP2 WAN links, despite their different available capacities.

1. Application Policy Hit Count Behavior

No known PR or KB matching this specific behavior was identified during the investigation.

The SSR application policy hit count increments when a new session matches the policy. It does not increment for every packet or byte transmitted over an existing session.

During troubleshooting:

• Active sessions were present.
• Live packet captures confirmed that traffic was flowing as expected.
• After clearing the application's sessions, the hit count increased by four, corresponding to four new sessions, and then remained constant while traffic continued.

This behavior is consistent with long-lived NFS sessions. A stable hit count does not necessarily indicate that traffic has stopped. To verify active traffic, check the application's sessions and throughput.

2. Long-Lived NFS Sessions

NFS sessions can remain active for extended periods. The SSR removes sessions based on the configured inactivity timeout; sessions that continue exchanging packets, including keepalives or lease-related traffic, may not become idle long enough to expire.

The session lifetime depends on the NFS client and server configuration, including lease settings, TCP keepalives, and mount options.

Recommendation: Engage the Rubrik application team to confirm the expected session lifetime and relevant NFS/TCP settings. Historical session behavior can also be reviewed using the Premium Analytics dashboard, where the required data is available.

Reference: NFSv4 Protocol Specification

3. WAN Link Utilization and Traffic Steering

During backup transfers, the following link utilization was observed:

WAN Link Observed Throughput

ISP2 (ge-0/0/4) Approximately 50 Mbps

ISP1 (ge-0/0/5) Approximately 350 Mbps burst; 250 Mbps sustained

The application object sed the traffic steering policy, which assigned equal weights to both WAN links. In contrast, another application object created for another backup used a traffic steering policy, with ISP1 as the primary link and ISP2 as the backup.

The equal-weight load-balancing configuration could place long-lived backup sessions on the lower-capacity ISP2 link, limiting the throughput of those sessions and contributing to the observed performance issues.

The following weight changes were recommended:

• ISP1: Cost/weight 3
• ISP2: Cost/weight 1

After applying the changes, the path application information was updated. However, existing sessions continued to use their established paths, so the new weights might not affect traffic until new sessions are created.

Recommended next steps:

1. Obtain approval from the application owner to clear the existing Rubrik-Backup sessions during a maintenance window when no critical transfers are running.
2. In Mist, navigate to Testing Tools → Applications → Sessions.
3. Select the application and delete the selected sessions.
4. Verify that new sessions use the updated traffic-steering weights and check throughput on both WAN links.
5. If TCL is intended only for failover, consider using traffic steering policy, with ISP2 as primary and ISP1 as backup.

4. Recommended SSR130 Release

To identify the currently recommended firmware release for the SSR130, navigate to:

WAN Edges → Node0/Node1 → Utilities → Upgrade Firmware

Review the versions listed in the Suggested dropdown. These represent the recommended releases presented by the Mist portal for the selected device.

The available screenshot was from an SSR1300 appliance. Verify the suggested version for the specific SSR130 device in the customer's organization before proceeding with an upgrade.

Conclusion

The investigation did not identify a confirmed application policy hit-count defect. The observed behavior is consistent with long-lived NFS sessions, where hit counts remain stable while existing sessions continue carrying traffic.

The low-throughput issue may be related to sessions being distributed across WAN links with different capacities. Applying appropriate traffic-steering weights and recreating sessions—with application-owner approval—will help validate whether the revised distribution resolves the performance issue.

The recommended firmware release should be verified directly in the Mist portal for the affected SSR130 device.