Backup and archive operations fail between SPS and NFS server (4383081)

When an NFS backup mount becomes stale (e.g. due to an intermediate firewall reboot), SPS continues to return success on subsequent calls without verifying that the mount is actually functional. This causes all subsequent backup and archive operations to fail until manual intervention or a full appliance reboot.

This persistent connection becomes problematic in environments where stateful intermediate devices such as firewalls sit between SPS and the NFS server. When such a device is rebooted or its session table is cleared, it drops the NFS session state. SPS, unaware of this, continues sending backup traffic over what it considers a valid connection. The middlebox silently drops the packets, and because the NFS mount on SPS never times out, this condition persists indefinitely without triggering any automatic recovery or alerting.

Symptoms include:

• Backup or archive jobs appearing to run but silently failing due to dropped packets
• NFS mount remaining in a stale state with no timeout
• Log entries such as:
  
  

Error performing backup; Could not create destination directory; 
destination='/mnt/platform/<uuid>', error='mkdir(): Permission denied'

STATUS

Change Request # 706819 was created to address this issue in a future release of SPS.

WORKAROUND

• Reboot the SPS appliance to clear the stale NFS session and restore backup functionality after a connectivity disruption.
  • There is currently no other supported workaround to force SPS to close and re-establish the NFS connection without a full appliance reboot.
• Ensure NFS server and network path stability to minimize the frequency of session drops in the interim.