I recently got a couple of SAS3 drives from a friend, they were reported as working, with the partition tables removed prior to me receiving them. I have a Supermicro Server with a SAS3 backplane with a Supermicro JBOD connected to it. Currently between the server and the JBOD I have 44 SATA disks being recognized and in use. I first inserted the SAS drives in the JBOD, the light blinks solid blue about once per second, different from the quick blue flash on all the drives actively in pools. The drives however were not appearing in the disks menu in the GUI. I then moved the drives to two open slots on the SAS3 backplane in the server, same LED behavior and still not showing up in the disk menu. I ran an lsblk -a in the shell and only the existing 44 drives are appearing.
Some googling indicates it could be due to the drives previously being used and so TrueNAS doesn’t want to recognize as a fresh ready to initialize disk. I don’t know what to do about this though since its not showing up in the lsblk output, there’s no way to wipe the drive, plus it was reported that all the partitions were already removed, but I can’t verify that myself. I don’t have any other machines with SAS connectors available. The hardware specs are:
Chassis is Supermicro SSG-6029P-E1CR12T (826BE1C4-R1K23LPB) X11DPH-T
HBA for Backplane Supermicro AOC-S3008L-L8E1-NI22 12G 8-Port SAS Host Bus Adapter
HBA for JBOD LSI SAS9207-8e 8-Port External HBA
JBOD - 45 bay supermicro (I don’t have the exact model, its a SAS2 backplane, 847 something)
I would think that even 520 byte sector, (or the 4K equiv), will show up with lsblk.
My own thought is that the disks may have power disable pins. (You don’t list these drive’s model…) This feature is more common with SAS drives than SATA drives. The test is to try and block off one of the drive’s first 3 pins. See below for some information:
You could also look up the disk model on the manufacturer’s web site and check out it’s manual. See if the manual lists power disable pins as a feature.
I believe this is going to be the answer. It is a Western Digital DC HC560, P/N: 0F38652 which corresponds to the SAS 20TB Secure Erase. The full Model is WUH722020BL5204. I dug through the datasheet for the drive, the 5204 at the end of the model, the 52 corresponds to 512e sector size, the 0 is Power Disable Pin 3 Support and the 4 is Secure Erase. So this seems like the likely cause. I’ll be trying to cover Pin 3 and will report back.
I figured this was going to be the answer, it made sense, and when I pulled the drives after they had sat in the chassis all night they were still cool to the touch, so even more evidence that they were not spinning up. I blocked pin 3 on the drives and they still were not being recognized. Worried that my tiny strip of label maker tape over the pin was not enough I also tried blocking off pins 1-3 entirely, the tape held much better but its still not being recognized. I tried inserting one of the drives, with the bay above it empty, so I could hold my fingers on the drive as it seated, it does feel like it might be spinning up, but was hard to tell. Either way its still not recognized in lsblk or the GUI. Could it have something to do with how the drives had the partitions removed? They were “tested verified working” before they were given to me and I have no reason to believe otherwise.
edit: by tested verified working I mean they were working to the point that the drives/partitions were recognized to the point of being able to delete the partitions before the drive was given to me.
If the drive appears to be spinning up, my next thought is that the SAS controller is not setup with IT firmware, (Integrated Target, as opposed to IR, Integrated RAID firmware).
But, that makes less sense as you have other drives showing up…
Yea they are both HBA’s, not RAID controllers, and I’ve tried using two different backplanes connected to two different HBA’s. One backplane/HBA is SAS2 the other is SAS3.
