SAS hard drives are not recognized after removing RAID and IT Mode Flash – help needed

[hpoperator]

hello dear Community,

I am currently working on a problem with several SAS hard drives that were previously used in a hardware RAID. I have reflashed an LSI 9211-8i controller several times to IT mode following these instructions: [Guide] Updating your LSI SAS Controller with a UEFI Motherboard which worked without any problems—the controller is recognized correctly by the system everywhere. The actual problem is that although the SAS disks appear, I cannot access them, regardless of whether I use SystemRescue, Ubuntu Live, or WindowsPE.

Under SystemRescue, I tried various commands, including lsblk, lsscsi, sg_map, and sg_inq, but although the disks are listed, I am still unable to access them, as mentioned above. I then installed OpenSeaChest-Basic under Ubuntu Live in the hope of being able to change the sector size or at least address the disks, but again, I was unable to access them. The same was true under WindowsPE with Disk Management and Diskpart. I also tested another PCIe slot, another 9211-8i (I have 4 of them, with a total of 24x 10TB SAS hard drives, and I successfully flashed all cards in IT mode), and I installed a 9211-8i in a completely different system, but that didn't change anything either.

When I plug all the 9211-8i cards into the PCIe slots at the same time, they are displayed correctly in the LSI BIOS. It also shows how many and which SAS disks are connected to which card.

For the sake of simplicity, I have now decided to leave only one 9211-8i plugged in and unplugged the rest.

The SAS hard drives are HGST HUH721010AL4200 with revision level: A923.

My current assumption is that the disks may still have some proprietary RAID metadata or are locked and therefore not recognized. There is also the possibility that the disks are currently formatted with 520 bytes or 528 bytes and therefore cannot be accessed. It may be necessary to perform a low-level reset, but I am unsure which tool would be suitable for this. OpenSeaChest and other standard programs have not been helpful so far.

If anyone has any ideas on how I can reinitialize and reformat the drives, I would be grateful for any tips. Perhaps someone has had similar problems with SAS disks after removing a hardware RAID and switching to an IT mode controller.

beste regards!

 

[EasyRhino]

you seem to be attemptin to reformat with sg_format to 512B, which would be the fix if it was a 520 byte block problem.

Let's start with the most common thing. 3.3v power. Thes generation of HGST drive will probably need 3.3v power disabled so the drives don't get stuck in reset loop. There are a variety of ways (kaptan tape etc), but my favorite is when using a cheap SATA power splitter cable to just cut the 3.3v wire.

 

[hpoperator]

you seem to be attemptin to reformat with sg_format to 512B, which would be the fix if it was a 520 byte block problem.

Let's start with the most common thing. 3.3v power. Thes generation of HGST drive will probably need 3.3v power disabled so the drives don't get stuck in reset loop. There are a variety of ways (kaptan tape etc), but my favorite is when using a cheap SATA power splitter cable to just cut the 3.3v wire.

Thank you very much for your reply!

Just to clarify: this server has been in use for years with exactly this configuration. Does the change you suggest to the 3.3V supply and the problem with the reset loop relate to general operation or only to the formatting process?

Regardless of this situation, the fact is that I cannot initiate the formatting process at all. I even tried sedutil because I briefly thought that the disks might be encrypted, but the sedutil commands did not get me any further either.
It is also mentioned elsewhere that the PSID should be located on the label of the hard drive and must have 30 characters. I do have an alphanumeric string on the label, but it only has 20 characters. Furthermore, Sedutil does not display any encryption anyway.

I just don't know what to do anymore. I can see the hard drives, but I still can't access them.

sedutil-cli --scan also does not show me any encryption; it has the status “No” for Opal.

Regarding my research into whether the hard drive might have been formatted with a different sector size, 520 or 528, the output of
blockdev --gets /dev/sdb

512

So apparently, the disk is already formatted in a compatible sector size.

I'm simply at a loss.

 

[nexox]

All I can guess is the HBA is doing something, can you browse around the LSI BIOS interface to see properties or settings for individual disks in case there's something there? Not that I recall what the BIOS options could be for that era of HBA, but I can't see how any other component would cause this combination of responses.

 

[kapone]

this server has been in use for years with exactly this configuration

Not quite. You reflashed the card(s). It looks like something is not right with your flashing.

Any chance you have an unflashed (i.e. you didn't try to flash the card to anything) card you can test with? This series of cards is dirt cheap nowadays. It may be worth it to get another card (and NOT flash it) just for troubleshooting.

 

[EasyRhino]

you tried a completely different 9211-8i that wasn't the one you reflashed, and same problem?

Okay I would think there are some magic bits (a superblock?) left on the disk from the hardware raid configuration.

probably dd or wipefs can do it in linux. in windows it would be diskpart -> clean I think.

How to wipe RAID information - Data Storage, Backup & Recovery - Spiceworks Community