VCS NFS/CNFS Agents Show as "Dead" Under systemd Despite Normal Operation

Problem Statement

Administrators monitoring VCS-managed NFS or CNFS agents under systemd may observe that the following systemd units report as “dead” or “inactive” even though NFS functionality remains unaffected:

• vcs_statd.service
• vcs_mountd.service
• vcs_idmapd.service

Example Output:

# systemctl status vcs_statd

● vcs_statd.service - VCS Managed rpc.statd

   Loaded: loaded (/etc/systemd/system/vcs_statd.service; enabled)

   Active: inactive (dead)

Despite this appearance, the underlying NFS services (e.g., rpc.statd, rpc.mountd, rpc.idmapd) continue to function correctly.

Environment / Context

• VCS NFS/CNFS agents running on RHEL 7.x, 8.x, 9.x or compatible systemd-based Linux distributions.
• Systemd is used for service management.

Cause / Root Cause

The current design of the VCS NFS/CNFS agent scripts manages the NFS-related processes (rpc.statd, rpc.mountd, and rpc.idmapd) directly.

• These processes are invoked under the main vcs.service unit, rather than their individual vcs_*.service units.
• As a result, systemd may report these sub-services as “dead” or “inactive”.
• This is expected behavior with the current VCS agent implementation and does not indicate a fault or service outage.

Process Mapping Example:

# ps -o pid,args,unit -C rpc.statd

    PID COMMAND                     UNIT

 754150 /usr/sbin/rpc.statd         vcs_statd.service

# ps -o pid,args,unit -C rpc.idmapd

    PID COMMAND                     UNIT

 754156 /usr/sbin/rpc.idmapd        vcs.service

• rpc.statd is associated with vcs_statd.service due to VCS scripts using systemd methods for process management.
• No legacy (non-systemd) restart or kill methods are invoked.

Resolution / Solution

There is no functional impact to service availability.

• NFS and CNFS operations continue to run normally.
• Client connectivity is not affected.

Verification Steps:

• Use the following commands to confirm process state association:
  • systemctl status <PID>
  • ps -o pid,args,unit -C <process>

These commands confirm which systemd unit each process is mapped to.

Symptoms / Expected Behavior

• Systemd units for VCS NFS/CNFS agents may show as “inactive” or “dead”.
• NFS/CNFS services remain operational.
• No impact to client access or data availability.

Additional Information / References

• RHEL systemd documentation:
  • man systemctl
  • man ps
• VCS 7.x Agent Pack documentation

Future Enhancement

Veritas Engineering is actively working to update the NFS/CNFS agent design to improve systemd integration and better align unit reporting with standard systemd practices.

• This change will be included in a future agent release, not as a hotfix.

Applies to:

• Veritas Cluster Server (VCS) for Linux
• VCS Agents for NFS and CNFS (7.4.2 and greater)
• RHEL 7.x, 8.x, 9.x or compatible systemd-based distributions