S2600CP2 - On losing PSU power, it won't post, except after a CMOS flash, even on a clean shutdown from OS - Feature or known issues and how to change if its a feature?

[ROles1]

S2600CP2 - On losing PSU power, it won't post, except after a CMOS flash, even on a clean shutdown from OS - Feature or known issues and how to change if its a feature?

I have two S2600CP2 boards, identical setups with 2670s.

They both post when have cleared CMOS via the jumper and a boot-and-wait.

They have both had CentOS installled and report all fine with CPUs and Memory and I ran "stress -c 256 -m 256" for 5 minutes and ran i7z to monitor temps and they are fine.

Whole board has adequate power and cooling.

Both boards have fresh CR2032 batteries and newest BIOS versions.

Issue is:

- On both boards, because I am still experimenting with a HW setup, they are not cased yet but on an anti-static testbench and everything there works fine and they are only touched with a grounding wire and as before work fine when "freshly CMOS booted" OR they have continous power from the PSU on reboots.

- When I then turn them off in a safe position like on a prompt or not in the middle of something like the BIOS or so, I get a quiet dual green light "failure to post" but with no ambers on.

Interrim, testbench solution,

- I then turn it off again on the PSU button. I wait 5 sec for everything to be safely turned off, before flipping the PSU button on again (every PSU on-off-on)

- Move the CMOS jumper to "clear" (2-3) and boot again to a "quiet failpost" (on purpose because it just clears the CMOS and resets the BIOS), where the green light on the CPU2 side of the board next to the blinking blue light goes out for 10 sec when the blue blink goes away and then comes back on (indicating CMOS cleared) and I turn off the PSU again.

- Then turn it on and it posts and gets into the BIOS or OS just fine and resumes operations. I usually go by the BIOS and check for boot target no 1 is the SSD for the OS though.

Conclusion:

- While there is a solution to this problem (move the jumper and triple boot over the course of about 90-120 seconds) or mount it in a case and never turn off the PSU (which I intend to do when done testing), ...

- ... I would still like to know if there is a way for the BIOS to stop this behavior and just post or for it to reset itself if it posts badly or SOMETHING to avoid busting fingertips on moving that jumper so much and boot those times?

- Is it intended serverroom behavior as it could (in there) be interpretted as a PSU AND an UPS problem, ie catastrophe (hurricane took the roof off, fire or grand theft) so it just shuts down?

- Any software fix or is there a BIOS fix for it or should I just get it case mounted and never turn off the PSU?

- Is there a way to (without buying a whole case) buy a power on/off/reset button to not have to touch the PSU on a testbench?

tia.

[ROles1 · Accepted solution]

Solved. I could only boot via PSU straight to BIOS with a clean CMOS.

I had no case (at that time, due to mounting and remounting CPUs till I was grey in the face from TIM). Once the case was there and there is a power button wire hooked to the Power, Reset and HDD light panel-of-pins, then you can let it boot silently to the 2 green lights (on a full unplug of the PSU) and then hit the power button and it just boots, np.

So the solution is to case the board or add a power button wire hooked up to the MB.