TCA CSAR upload fails with 'VNFD parse service failure' due to full /logs partition

Products

• VMware Telco Cloud Automation

Issue/Introduction

When attempting to upload a CSAR file to TCA, the following symptoms occur:

• The upload fails in the TCA UI or via API with the error: VNFD parse service failure.
• The system returns a java.net.SocketException: Connection reset error.
• The issue is persistent across different CSAR files but may appear intermittent depending on log rotation cycles.

Environment

3.2

Cause

The root cause is Management Domain resource exhaustion on the TCA Manager appliance. Specifically, the /logs partition (typically /dev/mapper/vg_logs-lv_logs) reaches 100% capacity.

When this partition is full, the Java-based tca-catalog and vnfd-parsing services cannot write the temporary metadata or transaction logs required to parse and validate the incoming CSAR/VNFD payload. This leads to an ungraceful termination of the TCP connection, manifested as a socket reset.

Resolution

To resolve this issue, you must reclaim space on the TCA Manager appliance log partition.

Verify Disk Space

1. Log in to the TCA Manager appliance via SSH as admin and switch to the root user.
2. Review this command before running it.
   

   df -h /logs

3.  Confirm if the usage for /dev/mapper/vg_logs-lv_logs is at or near 100%.

Reclaim Space

1. Navigate to the kubelet service retained logs directory
   

   cd /logs/retained-logs/kubelet.service
   

2.  Delete old log files in this directory to reclaim immediate space.
3. Run the following command to identify and delete files larger than 40MB or old compressed archives across the /logs partition.
   

   find /logs -type f \( -size +40M -o -iname "*.tar.gz" \) -delete

Validation

1. Verify the disk space again using df -h /logs to ensure sufficient free space is available.
2. Attempt to upload the CSAR file again via the TCA UI or API.