Proxmox VE 9.2.4 (MS-01) - pvestatd permanently blocks in clone()/copy_process() (D state), causing VM/LXC status to become '?' until restarted

[RejectedGranny]

Summary​

pvestatd periodically becomes permanently blocked, causing all VM/LXC status in the web UI to become ?. Guests continue running normally and the web interface remains responsive.

Restarting pvestatd immediately restores the UI until the issue occurs again.

This is a standalone node (no cluster).

Environment​

• Proxmox VE: 9.2.4
• Kernel: 6.14.11-4-pve
• Hardware: Minisforum MS-01
• Storage:
  • local
  • local-lvm
  • NFS storage
  • Proxmox Backup Server datastore
• No ZFS

Symptoms​

• VM/LXC status changes to ?
• Guests continue running normally
• Web UI remains accessible
• systemctl restart pvestatd immediately restores status

Earlier occurrences also showed:

Main process exited, code=killed, status=11/SEGV

However, the current issue is different: the process does not exit and instead remains permanently blocked.

Investigation​

pvestatd enters uninterruptible sleep:

State: D (disk sleep)
Kernel wait channel:

copy_process
Kernel stack:

copy_process
kernel_clone
__do_sys_clone
__x64_sys_clone
x64_sys_call
do_syscall_64
entry_SYSCALL_64_after_hwframe
/proc/<pid>/syscall reports syscall 56 (clone).

The stack remains identical indefinitely.

The process cannot be terminated normally.

Interestingly:

• /proc is accessible.
• busctl list works.
• NFS storage remains responsive (ls /mnt/pve/share completes immediately).

However:

• systemctl status pvestatd hangs.
• systemctl is-active pvestatd hangs.
• journalctl -u pvestatd hangs.
• ps -ef hangs.

This appears to be pvestatd blocking during clone() rather than during storage access.

I can reliably recover by running:

systemctl restart pvestatd
until the issue reoccurs.

Please let me know if additional diagnostics (hung task dump, sysrq output, crash dump, etc.) would be useful.

Thank you!