Hi

 

I have the problem that several discs are no longer recognized and I can not find the cause. Maybe someone can interpret the logs better than I can and put me on the right track.

 

# Setup/Background

The setup as it is currently worked on my old motherboard (without GPU). I have two parity disks, as well as two caches - one pool for appdata and VMs (SSD) and a normal cache (NVMe) - and I have a SAS card in IT mode for additional SATA ports.

 

# History

I moved to a bigger case and in this context I also changed my mainboard and processor. After the move everything worked perfectly. So I decided to do another planned upgrade (step by step) and installed a Nvidia GPU. That wasn't quite so trivial, and I booted and shut down the system about 20 times in that context. It's quite possible that something broke in the process.

 

# Problems

Two of my data disks are currently missing completely. The cache pool (Vm) was also missing at first, but now one SSD is back - the second one is still missing.

 

# I have already tried this...

I changed the power and SATA cables and reconnected them. Even if I plug them directly onto the motherboard or the SAS card or removing the GPU - it makes no difference.

 

I have also thought of a power problem, but according to various online calculators, I should be fine with my 650 watt power supply.

 

Maybe someone has more ideas?

 

Your help is much appreciated.

Best,

0x3

 

Edit: removed Diagnostics

 

Edited February 27, 20224 yr by 0x3

Make sure the missing devices are all connected to the onboard SATA controller and see if they are detected by the board BIOS.

That was a good tip, thank you JorgeB.

 

I removed all the cards and the GPU and only connected the 3 (4) missing discs directly. Works. I was able to read the SMART status in BIOS.

 

Then I did the following: screwed the M.2 onto a PCIe card so that all onboard SATA slots are available (in theory). Since I had too little space, I connected another PCIe SATA card and connected the remaining disks there. That was a bit rushed. It doesn't seem to make any difference - 5 disks are now missing.

 

Somehow there seems to be a limitation on disks - no matter how I connect them (SAS or SATA card). I don't know what this limitation is and why it worked before.

 

Do you have an idea? Is this a BIOS setting or known limitation on some motherboards?

10 minutes ago, 0x3 said:

Then I did the following: screwed the M.2 onto a PCIe card so that all onboard SATA slots are available (in theory). Since I had too little space, I connected another PCIe SATA card and connected the remaining disks there. That was a bit rushed. It doesn't seem to make any difference - 5 disks are now missing.

Not really following, what disks are missing? The ones on the PCIe controller? Usually boards share some lanes, so using for example an M.2 slot might disable a PCIe slot, also using M.2 SATA devices will generally disable some SATA port(s), consult the board manual.

It looks pretty random, so about 50:50. There are missing disks that are directly connected and also some from the PCIe card. According to the manual, a SATA port is disabled when the M.2 is in SATA mode. But I have removed the M.2 as a test. So seems to be a different issue.

Edited February 27, 20224 yr by 0x3

Sounds to me like a power/connection problem, if the same devices sometimes work, sometimes don't, problem is probably on the power side.

It is indeed very strange. In further tests, all disks on the PCIe SATA card were not available. I then removed the SATA card and reconnected the SAS. Now suddenly all disks are online. The question is how stable this runs.

 

Do you have any ideas to investigate a possible power issue?

Try with a different PSU if available.

Unfortunately no. But maybe I'll just buy one so I have a reference.

 

I really appreciate your help - thank you very much!

I had several problems at once:

 

1. The power supply - I could not confirm this clearly, but it is very likely the case.
2. Empty CMOS batteire, which caused the settings in the BIOS to reset when the power was interrupted. Every time I rebuilt hardware components, I hit the power switch.
3. RAM was 2400 MHz instead of 2666 MHz - this was supported by the mainboard, but not by the CPU according to the specification. This resulted in occasionally getting an error message and not booting.

Currently the setup is running stable and I will now try to connect the GPU.