This model isn't available on SAS drives.
This model isn't available on SAS drives.
Drives unexpectedly disconnect, even when cold, must there be a configuration problem? I acquired my server a year prior and promptly purchased several new drives. I set everything up in RAID and it operated smoothly on Ubuntu. After a few weeks, I powered it down and left it idle while working on assembling a server rack for all components. Circumstances shifted, but I completed the build, installed Proxmox, and discovered that many drives weren’t functioning properly. I’m aware drives can fail, yet only certain ones are affected. Specifically, all the drives I bought last year. Yes, you’re right—they were used, so they should fail eventually. All of them? At once? Even the cold spares stored away in a drawer? It seems improbable, so I suspect an error in setup or hardware damage. When trying to access them, the system reports an error: "The requested command cannot be completed, as the specified device does not exist." Working drives fit any slot, whereas failing ones don’t, indicating it’s not a backplane issue.
Server Specifications: Dell R730XD, BIOS v2.5.5 v2.19.0, 384GB ECC DDR4 1866MHz, Dual Xeon E5-2697v3 @ 2.6GHz, Dell PERC H730P v2.50 v25.5.9 RAID card.
Working Drives: DELL ST1200MM0088 1.2TB 10K SAS (x6)
Failing Drives: HGST HUC101212CSS600 1.2TB 10K SAS (x18)
Troubleshooting Steps: Installed only one HGST drive, moved DELL to the previous failing HGST slot. Both power supplies were connected. Defaulted BIOS settings. Cleared RAID controller config. Tried HBA mode and mass update via Lifecycle, but Dell stopped providing downloads.dell.com. Pulled packages from support, updated BIOS to v2.19.0 with Cloud Ninjas help. Attempted iDRAC update via IRAC using .EXE files; encountered errors about file types. Learned from a PowerEdge Server KB that the path must be set to catalog. Failed connection to network share (SWC0066). NIC was configured for DHCP, so moved on. When updating iDRAC via Update Package, it reported "Unable to extract payloads" and "Unable to verify signature." Found a firmware update for specific drives in v2.86, then PERC v25.5.9, but encountered errors about missing firmware image. After a full reboot, the issue persisted.
I tried restarting the Lifecycle Controller, using HTTPS, and eventually managed some updates, though problems remained. Ultimately, all my HGST drives stopped being recognized, even after extensive fixes, and nothing resolved the issue. TL;DR: After installing a complete RAID5 setup in Ubuntu, I shut down the system for a year. Now, all my HGST drives aren’t detected, despite full updates and no apparent hardware failure. It’s puzzling—especially since they were purchased used and should have failed eventually."
New Development: Removing and reinserting drives affects which ones appear in the virtual disk setup within iDRAC, limited to specific models. For instance, with all x6 DELL and x18 HGST drives installed, swapping a DELL drive shows only its type while hiding others. Similarly, swapping an HGST drive reveals just that model. Note: Creating a virtual disk using solely HGST drives doesn't work.
Following several power cycles and configuration adjustments—such as activating HBA mode—the HGST drives are now displaying information in iDRAC's Physical Disks, a behavior that hasn't occurred before and I'm unsure why. The firmware update appears to be linked to this change, but I'm still facing difficulties installing it. Based on the Reddit discussion, an incorrect block size might be the problem. Using sg3_utils, as demonstrated in the guide, could help ensure proper recognition by the controller. Note: After installing Proxmox and opening the Shell window, sg_scan -i only shows the DELL drives, not the HGST drives that are blocked.
After thorough investigation and verification, I’m convinced the problem lies with the HGST block size not matching the PERC H730p requirements. I’d likely need to link these drives to another system or HBA, then reformat them using sg3_utils to a 512-block configuration. Until then, they’re essentially unusable. I’m puzzled by how the drives changed their format after being powered off for a year—they were functioning previously. It’s possible that adjusting controller settings during reimaging or server setup altered their settings. Ultimately, I’m leaning toward using the PERC H730p in HBA mode. With the latest firmware, it performs well with Proxmox and offers solid performance. Although not a perfect 100% pass-through, it’s sufficient for now. It’s frustrating that the card blocks drives it thinks incompatible, even when running in HBA mode. I’d like to switch the card into IT Mode, but I don’t have the correct firmware or procedure for the H730p (only H310/H710/H810 are available), and this doesn’t guarantee a true pass-through.
They are being processed using the RAID card in HBA mode with ZFS on Proxmox.
Sure thing. It’s been a bit confusing how long it took me to get everything sorted, but I managed to set it up using my NDS-4600 over SFP+ to an LSI HBA in the R730XD. I transferred all the drives, reformatted them with the specified command, loaded them back in, and they functioned right away. It’s frustrating that it took so long to see the solution work clearly. I even tried a GODO SAS Dock, which didn’t work and came with a damaged 3.0 Micro B, so I’m planning to return that. I hope this thread helps someone else later. Let me know if you have any questions.