I don't know why, but me giving it another try with an HBA card is turning into a nightmare.

I am trying to isolate the issue. Somehow, disks don't mount / cannot be mounted / cannot be added to the array. I am struggling to find the logic. I thought it was a heat issue when attaching more than four disks to the HBA. Got another HBA to connect 4 each. But no improvement. The second one is also meant to be more energy efficient, but it seems equally unstable (or even worse).

Maybe the easiest is to share a diagnostic log. Maybe you can see anything and make some sense what's going so wrong? Thanks for your help!

The log is with only one HBA connected. With only 4 disks connected to the HBA. None of them being identified and none of them possible to mount.

tower-diagnostics-20251002-2208.zip

There are what look like power/connection issues with multiple disks; since they affect multiple controllers, most likely a power problem:

Oct  2 22:09:05 Tower kernel: sd 9:0:1:0: [sdg] Attached SCSI disk
Oct  2 22:09:05 Tower kernel: sd 9:0:1:0: [sdg] Synchronizing SCSI cache
Oct  2 22:09:05 Tower kernel: sd 9:0:1:0: [sdg] Synchronize Cache(10) failed: Result: hostbyte=0x01 driverbyte=DRIVER_OK
Oct  2 22:09:05 Tower kernel: mpt3sas_cm0: mpt3sas_transport_port_remove: removed: sas_addr(0x300605b00d3a84a6)
Oct  2 22:09:05 Tower kernel: mpt3sas_cm0: removing handle(0x0021), sas_addr(0x300605b00d3a84a6)
Oct  2 22:09:05 Tower kernel: mpt3sas_cm0: enclosure logical id(0x300605b00d0984a0), slot(7)
Oct  2 22:09:05 Tower kernel: mpt3sas_cm0: enclosure level(0x0000), connector name( C1  )
Oct  2 22:09:09 Tower kernel: ata7: link is slow to respond, please be patient (ready=0)
Oct  2 22:09:14 Tower kernel: ata7: SATA link down (SStatus 0 SControl 300)
Oct  2 22:09:14 Tower kernel: ata7: limiting SATA link speed to <unknown>
Oct  2 22:09:20 Tower kernel: ata7: link is slow to respond, please be patient (ready=0)
Oct  2 22:09:27 Tower kernel: ata5: SATA link down (SStatus 0 SControl 300)
Oct  2 22:09:28 Tower kernel: ata7: SATA link up 6.0 Gbps (SStatus 133 SControl 3F0)
Oct  2 22:09:29 Tower kernel: ata7.00: configured for UDMA/133
Oct  2 22:09:29 Tower kernel: ata7: limiting SATA link speed to 3.0 Gbps
Oct  2 22:09:29 Tower kernel: ata7: SATA link up 3.0 Gbps (SStatus 123 SControl 320)
Oct  2 22:09:29 Tower kernel: ata7.00: configured for UDMA/133
Oct  2 22:09:32 Tower kernel: ata6.00: exception Emask 0x50 SAct 0x80000000 SErr 0x4090800 action 0xe frozen
Oct  2 22:09:32 Tower kernel: ata6.00: irq_stat 0x00400040, connection status changed
Oct  2 22:09:32 Tower kernel: ata6: SError: { HostInt PHYRdyChg 10B8B DevExch }
Oct  2 22:09:32 Tower kernel: ata6.00: failed command: READ FPDMA QUEUED
Oct  2 22:09:32 Tower kernel: ata6.00: cmd 60/08:f8:30:2a:9b/00:00:a6:0c:00/40 tag 31 ncq dma 4096 in
Oct  2 22:09:32 Tower kernel:         res 40/00:01:01:4f:c2/00:00:00:00:00/00 Emask 0x50 (ATA bus error)
Oct  2 22:09:32 Tower kernel: ata6.00: status: { DRDY }
Oct  2 22:09:32 Tower kernel: ata6: hard resetting link

Thanks. When you say power problem, you referring to the PSU?

Yes, or power splitters, etc

This reads like a hard to crack problem... Any idea how to get ahead to trouble-shoot?

I noticed things do seem to work well as long as I use only a few disks. Once I start using more than 4-6, challenges pop up.

When starting the server, all fans go on with a big boost. Then everything turns off again. And then everything starts again (without the "boost") and things properly start.

I have a solid PSU (NZXT C750), which isn't too old. So, I don't suspect that I am under-powered per se and also the PSU should still have life.

Edited October 2, 2025Oct 2 by steve1977

Ideally no more than 4 disks per PSU cable

I don't think I have this, but let me look into the cabling. I don't even have crazily many disks (8 with HBA, 5 with mobo, 2 nvme), so 750W feels plentiful? I also removed my GPU (though is old and shouldn't draw too much either), but this didn't bring any change.

Let me try whether it works when I run 6 disks from the HBA. I still somewhat have the hunch that my HBA card is the issue...

Here we go with another diagnostic file. Trying different combos how to connect to mobo and HBA to rule out it's an HBA (overheating) issue. If it's the PSU, it is far easier to solve. If it is a power cabling or a HBA overheating issue, far more trouble...

tower-diagnostics-20251003-0955.zip

Nothing in logs so far, but of course, too many missing disks to start the array so no real activity.

One purchase of a 1,650W PSU later, the issue seems resolved ;-)