I figured out how to get those set by shell.
To check the current “kernel_extra_options” config, mine was just "":
midclt call system.advanced.config | grep kernel_extra_options
To check the currently applied default_ps_max_latency_us, mine was originally 1000:
cat /sys/module/nvme_core/parameters/default_ps_max_latency_us
To apply @Beelink CS-George ’s suggestion:
midclt call system.advanced.update '{"kernel_extra_options":"nvme_core.default_ps_max_latency_us=0 pcie_aspm=off pcie_port_pm=off"}'
Then rebooted.
Checked that default_ps_max_latency_us was now 0:
cat /sys/module/nvme_core/parameters/default_ps_max_latency_us
After testing reboots and full shutdown cycles, it still did not solve the problem. Rebooting still loses the drives every time. Getting back into TrueNas showed missing pool and missing disks because BIOS still doesn’t see them on reboot. Full shut down is still needed to detect all 6 drives again.
I reverted @Beelink CS-George ’s suggestion with:
midclt call system.advanced.update '{"kernel_extra_options":""}'