NSX Upgrade Fails with GATEWAY_TIMEOUT and Host /tmp Partition Errors

Products

• VMware NSX

Issue/Introduction

When upgrading from NSX 4.2.2.1.0 to 4.2.3.2.0, the upgrade process fails during the pre-check or host upgrade phase with the following errors in the NSX Manager UI or upgrade-coordinator.log:

• Error 1 (NAPP): Failed to execute Check for NSX Application Platform version for upgrade. [UC] Error in rest call. url= /napp/api/v1/platform/monitor/platform/status , method= GET , response= upstream request timeout , error= GATEWAY_TIMEOUT:upstream request timeout.

• Error 2 (Host): Upgrade failed: Error while checking free space on tmp partition on host. Please check the configuration state of host and ensure that services are operating normally.

Environment

VMware NSX

Cause

 

• NAPP Timeout: The Upgrade Coordinator (UC) fails to receive a timely response from the NSX Application Platform's status API. This is often due to resource contention or service timeouts within the Kubernetes-based NAPP deployment during the upgrade transition.

• Host /tmp Full: ESXi hosts may have accumulated large log files, stale VIB remnants, or temporary upgrade bundles in the /tmp RAM disk, preventing the installation of new NSX components.

 

Resolution

To  successfully complete the NSX upgrade, you must manually decouple the problematic hosts from the automated workflow, proceed with the NSX Manager/Infrastructure upgrade, and then re-introduce the hosts.

Step 1: Unprepare the Affected ESXi Hosts

Before continuing the upgrade, remove the NSX components from the hosts reporting the /tmp space error.

1. In NSX Manager, navigate to System > Fabric > Nodes > Host Transport Nodes.

2. Select the affected ESXi hosts and click Remove NSX.

   • Note: Ensure VMs are migrated (vMotion) off these hosts if you do not have a redundant N-VDS configuration.

3. Confirm the "Unprep" process completes and the host status returns to Not Configured.

Step 2: Clear Host /tmp Space (Manual Cleanup)

On each affected ESXi host, perform a manual cleanup to ensure the partition is ready for re-preparation.

1. Log in to the ESXi host via SSH.

2. Identify large files using: ls -Slh /tmp | head -n 10

3. Remove known temporary upgrade logs or stale bundles:

   • rm -rf /tmp/vmware-root/*

   • rm -rf /tmp/nsx-update/*

4. Verify space with vdf -h.

Step 3: Continue the NSX Manager Upgrade

With the problematic hosts removed, the Upgrade Coordinator will no longer be blocked by their local partition errors.

1. Return to the Upgrade dashboard in NSX Manager.

2. Click Retry or Continue to proceed with the Management Plane and Control Plane upgrades.

3. If the NAPP GATEWAY_TIMEOUT persists, restart the NAPP status service or skip the NAPP check if the platform is not actively in use.

Step 4: Re-prepare the ESXi Hosts

Once the NSX Managers are successfully upgraded to version 4.2.3.2.0:

1. Navigate back to System > Fabric > Nodes > Host Transport Nodes.

2. Select the hosts you unprepared in Step 1.

3. Click Configure NSX (or re-apply the Transport Node Profile).

4. NSX will now install the version-compatible VIBs onto the cleaned /tmp partition.

Additional Information

NSX Upgrade Pre-check fails with the error message "Failed to execute Check for NSX Application Platform version for upgrade

Failed to execute Check for NSX Application Platform version for upgrade" during NSX Upgrade pre-checks