SOLVED Supermicro X10SL7-F and SAS Drives not found

[Mike Fuller]

So this is my first post, so please excuse my ignorance. Little about me - I've been putting together desktop PC rigs from scratch since the 1990's - but the server class hardware is fairly new. So here is the new NAS rig I threw together for the home network based on some recommendations I saw these very forums:

Mainboard: X10SL7-F
CPU: i3 Haswell
RAM: Crucial Unbuffered ECC DD3 1600 - 2x8 GB
PSU: EVGA Gold 650 watt
Chassis: Fractal Node 804
Boot Drive: Kingston 120GB SSD on SATA0
Planned ZFS2 Array: 6 x Seagate ST2000NM0014 2TB SAS 4Kn

All components are new. I put the box together, FreeNAS 9.10.2-U2 is on the metal and runs dandy. I did some flashing and after some learning how to use the UEFI shell, I have the the LSI 2308 controller running the latest 20.0.7.0 IR firmware and the AMI BIOS is running the 12/17/2015 version 3.0a. So the problem I have are the drives - the controller never "sees" them.

I have a benchtop PSU tester, and the power cables are energized. I am going sans backplane and just connecting the power and drives with these cables:

https://www.amazon.com/gp/product/B00HZS33U4/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1

The SFF8482 / SAS 29 Pin is plugged into each of the drive male sockets. The first free end of that cable I used the motherboard supplied SATA device cables to plug all six drives into the blue SAS headers on the motherboard. The SATA PSU power cables are coupled to the other free end. I e-mailed Supermicro and they said that should work although they haven't tested that exact cable from that specific manufacturer.

So...after setting every conceivable EFI / Legacy setting in the BIOS - the LSI2308 just will not see any of these drives. I cannot see the drives in FreeNAS either (no surprise). I can;t believe all six are dead. I am contemplating RMAing the board to Supermicro but I thought I would ask for any insight here first. Am I hooking everything up wrong? Anything I can test? Firmware I should try?

Thanks!
-Mike

 

[Spearfoot]

So this is my first post, so please excuse my ignorance. Little about me - I've been putting together desktop PC rigs from scratch since the 1990's - but the server class hardware is fairly new. So here is the new NAS rig I threw together for the home network based on some recommendations I saw these very forums:

Mainboard: X10SL7-F
CPU: i3 Haswell
RAM: Crucial Unbuffered ECC DD3 1600 - 2x8 GB
PSU: EVGA Gold 650 watt
Chassis: Fractal Node 804
Boot Drive: Kingston 120GB SSD on SATA0
Planned ZFS2 Array: 6 x Seagate ST2000NM0014 2TB SAS 4Kn

All components are new. I put the box together, FreeNAS 9.10.2-U2 is on the metal and runs dandy. I did some flashing and after some learning how to use the UEFI shell, I have the the LSI 2308 controller running the latest 20.0.7.0 IR firmware and the AMI BIOS is running the 12/17/2015 version 3.0a. So the problem I have are the drives - the controller never "sees" them.

I have a benchtop PSU tester, and the power cables are energized. I am going sans backplane and just connecting the power and drives with these cables:

https://www.amazon.com/gp/product/B00HZS33U4/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1

The SFF8482 / SAS 29 Pin is plugged into each of the drive male sockets. The first free end of that cable I used the motherboard supplied SATA device cables to plug all six drives into the blue SAS headers on the motherboard. The SATA PSU power cables are coupled to the other free end. I e-mailed Supermicro and they said that should work although they haven't tested that exact cable from that specific manufacturer.

So...after setting every conceivable EFI / Legacy setting in the BIOS - the LSI2308 just will not see any of these drives. I cannot see the drives in FreeNAS either (no surprise). I can;t believe all six are dead. I am contemplating RMAing the board to Supermicro but I thought I would ask for any insight here first. Am I hooking everything up wrong? Anything I can test? Firmware I should try?

Thanks!
-Mike

I don't think firmware is causing your problem, but you want to flash the motherboard LSI controller to IT mode for use with FreeNAS, not IR mode...

Why did you choose SAS drives? I use SATA drives with the same motherboard and get stellar performance. Plus SATA drives are cheaper... :)

 

[Mike Fuller]

I don't think firmware is causing your problem, but you want to flash the motherboard LSI controller to IT mode for use with FreeNAS, not IR mode...

Why did you choose SAS drives? I use SATA drives with the same motherboard and get stellar performance. Plus SATA drives are cheaper... :)

Well I got these at a price that was less than SATA at the same capacity / speed so figured what the heck.

I did some more testing - all the cables have power, and all SATA drives coupled to them spin up. (SAS connectors are compatible with SATA but not vice versa). The LSI2308 sees all SATA drives plugged in so all my data cables work good. Seems like there is nothing wrong with my mainboard or my cables. So now I am struggling with the possibility that all six of these Seagates are duds (possible but how likely?) or there is just something with the SATA power not giving enough amperage / incorrect voltage / polarity someplace. I may get legacy molex to SFF8482 cables and see if that solves it.

Any SAS geeks out there that can point me in the right direction before I go buying backplanes.?

 

[Spearfoot]

Well I got these at a price that was less than SATA at the same capacity / speed so figured what the heck.

I did some more testing - all the cables have power, and all SATA drives coupled to them spin up. (SAS connectors are compatible with SATA but not vice versa). The LSI2308 sees all SATA drives plugged in so all my data cables work good. Seems like there is nothing wrong with my mainboard or my cables. So now I am struggling with the possibility that all six of these Seagates are duds (possible but how likely?) or there is just something with the SATA power not giving enough amperage / incorrect voltage / polarity someplace. I may get legacy molex to SFF8482 cables and see if that solves it.

Any SAS geeks out there that can point me in the right direction before I go buying backplanes.?

Ummm... is it too late to exchange the SAS drives for SATA units? I hate to see you chase off on a backplane tangent.

I don't know much about SAS, beyond how to spell it... and seems like my daughters had plenty of it!

 

[Mike Fuller]

Where is the challenge in that? This is the perfect opportunity to study wiring diagrams and use the multimeter on the cabling ;)

Still its valid option and it seems like an overwhelming majority of users here just use SATA drives. I guess that's the thermonuclear option. I'd have to check my amazon and newegg receipts to see if I am still in the return window or how bad I'll get beat up on the overstocking fee.

 

[Spearfoot]

Where is the challenge in that? This is the perfect opportunity to study wiring diagrams and use the multimeter on the cabling ;)

Still its valid option and it seems like an overwhelming majority of users here just use SATA drives. I guess that's the thermonuclear option. I'd have to check my amazon and newegg receipts to see if I am still in the return window or how bad I'll get beat up on the overstocking fee.

Ah, you must be another electron pusher, like me! My weakness is vacuum tube guitar amplifiers.

Keep us posted about your progress. More data points in re: SAS drives would be useful.

Good luck!

 

[Arwen]

...I have the the LSI 2308 controller running the latest 20.0.7.0 IR firmware
...

Uh, you have the LSI Integrated RAID, (IR), software. So if the BIOS level LSI software,
(and potentially the OS level LSI RAID management tool), sees the drives, then you have
to use either a different LSI driver, (I don't remember it's name). Or you have reload the
LSI firmware to IT, (Intergrated Target).

 

[Mike Fuller]

So I tried three different cables / adapters similar to the one I linked. Some spin up the SATA drives, the SAS drives never spin up. Still research to see if SAS drives auto-spin or they need a command to spin-up. But it is odd that not all the connectors / adaptors are wired the same internally. Makes you wonder if all this stuff is made the same - or if there really is a more specific wiring I need.

 

[Mike Fuller]

Uh, you have the LSI Integrated RAID, (IR), software. So if the BIOS level LSI software,
(and potentially the OS level LSI RAID management tool), sees the drives, then you have
to use either a different LSI driver, (I don't remember it's name). Or you have reload the
LSI firmware to IT, (Intergrated Target).

This is true once I get the drives spinning and recognized by the controller. Then yes, the IT will let ZFS run more smooth. But any version of the 20.x.x.x firmware is fine because at this point all I want to do is get the physical drives registered on the LSI and I wanted to rule out any incompatibility with the firmware and this crop of SAS drives.

 

[Mike Fuller]

Missed my chance to return the drives to Amazon (looks like I got all 6 there in two different purchases.) Doubtful I can sell this to anyone on here because I have no idea how one would get them to work without an expensive backplane .

My last ditch effort would to buy these connectors:

http://www.pc-pitstop.com/sas_cables_adapters/I2917-05M.asp

But let the record show these have been tested:

1. https://www.amazon.com/gp/product/B00HZS33U4/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1

These work with SATA Drives (data and power check out) but not SAS Drives.

2. https://www.amazon.com/gp/product/B01MRPIFYG/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1

These did not work with my SATA nor my SAS Drives.

 

[Mike Fuller]

I have an older semi-modular PSU, I trimmed the plastic header on the SATA Power coupler and attached it directly to the SAS just to rule out these connector cables. It powers the SATA drives just fine (HDDs, SSDs, DVD-ROMs), but no spin up on any of the SAS drives. So there is more going on here since, according to the wiring diagrams, the voltage and pins on an SAS power male connector are the same as an SATA.

Guess I'll just RMA to Seagate and in the mean time go buy some WD Red Drives like a normal person and stop straying off the beaten path.

 

[Mike Fuller]

Update: Received my six RMAed drives from Seagate. Same issue persists. Nothing spins up when power is applied. So...its one of the following options:

1. Seagate has made a total hater out of me for life by giving me 12 junk drives in a row

2. Everyone who just hooks SAS drives directly to this Supermicro board and claims it works without a backplane is a liar.

3. There is some special trick or additional power supply needed to spin up SAS drives. It needs more juice than that provided by a conventional modular power supply? Does the polarity need to be reversed or something?

I re-checked the PSU against my testbed, all connectors are providing power. My other SATA HDDs, Optical Drives, SSD Drives all run fine on the PSU in my FreeNAS case and my test bed.

I am open to any suggestions. Otherwise, I think its time to cut my losses, toss my freenas server project in the garbage and stay with my existing Linux server. If someone lives local to me you can come to my house and take everything for a pittance. Way too much time, energy and money wasted to mess with this too much more...

 

[Jailer]

Did you read through the SAS primer thread?

https://forums.freenas.org/index.ph...-sas-sy-a-primer-on-basic-sas-and-sata.26145/

 

[Apollo]

Is there an option in the BIOS that state you are using SAS drives?

 

[Mike Fuller]

Jailer,

That thread is what encouraged me to go down the SAS path and give it a try. It all seemed very straightforward.

Apollo,

The BIOS does not even see the drives. All it sees is the SSD drive that I am using as the FreeNAS OS drive. Its like its not there. The SAS drives data cables are connected to the SAS controllers on the motherboard with a conventional SATA data cables that have the correct "end" to interface with an SAS drive as indicated in the primer and as recommended to me by Supermicro tech support.

 

[Mike Fuller]

Right now I am just trying to talk to someone who knows for sure if SAS drives spin up once the power is applied like any SATA drive. If I put the juice on the drive, even with no data cable connected - the platter *should* spin up. This has not been the case with any of these drives. I am almost tempted to buy the cheapest WD SAS drive I can get my hands on just to see how it behaves. It hard for me to rule out the drives without some other SAS device to test against. Truthfully, I got these drives at such a good price it was the same as the SATAs of the same capacity and vendor and figured - its enterprise right? Has to be better?...oh how hindsight is 20/20.

I double checked my Supermicro and Seagate documentation, both state that although the drives are designed for use in a backplane setup using cables is acceptable. Everything I have read says that other than the little plastic connectors on the ends of the data and power cables, SAS and SATA are identical in mechanical sense. The signal overt he data cable is a little different but otherwise it should be completely compatible.

Ordered a 18 dollar HP SAS 72GB drive...just for the heck of it to see if I can rule out Seagate...also opened a ticket with Seagate Tech Support to see if they have any insight.

 

[Arwen]

One thing to note about SAS disks, (and applies to parallel SCSI as well as Fibre Channel), is that they have
programable features. Meaning they could very well be configured from the factory not to spin up. Thus, allow
controller based staggered spin up.

That said, it's un-likely. Further, even spun down, they should still show up on a SAS controller.

It's good that your SAS controller sees SATA drives attached. That seems to eliminate many things. I did check
the specifications of your SAS drives, they seem to need just the plain 5volt & 12volt. SATA / SAS power
connectors can also supply 3.3volts, which is less common.

So, my current thought is that your SAS controller firmware does not like 4KB native disks. Seagate seems to
make several 2TB SAS models, 4KB native, 512B native, 512B emulated. Don't really know what's up. Hope
your test with the HP drive goes good. But you may want to check it's sectoring.

 

[Arwen]

New information. In another thread;

https://forums.freenas.org/index.ph...-sata-not-recognizing-hdds.54246/#post-376045

It has a link to this;

https://www.reddit.com/r/DataHoarder/comments/42hv7e/supermicro_x10sl7f_not_seeing_my_new_6tb_disk/
https://www.reddit.com/r/DataHoarder/comments/42hv7e/supermicro_x10sl7f_not_seeing_my_new_6tb_disk/
Hmm, give that a try.

Note that most drives that spin down, will spin up on access. So even if they don't spin up on, continue with testing to see if the BIOS
or FreeNAS sees the drives.

 

[Mike Fuller]

Hi Arwen,

Thanks for responding again. So as far as those threads go and their recommendations my comments and observations are these:

1. I tried both IR and IT firmware for 20.0.7.0 to no avail

2. I tried IR and IT firmware with EFI set and not Legacy to no avail. I am using UEFI to boot FreeNAS.

3. I tried seeing the drives in the Mainboard BIOS, The LSI BIOS, and in FreeNAS itself - no luck. FreeNAS just reports there are no drives attached (other than the OS SSD) to build a ZFS array (you can't use the OS SSD for the ZFS)

4. As far as the guy claiming his board came with a breakout cable - mine did not. And there is no place on the motherboard to connect anything other that a standard data cable (no mini SAS attachment points exist.

5. I confirmed the jumper for the LSI is enabled correctly, I switch the position for grins and on reboot the LSI SAS card is not even present in the BIOS. Switching the jumper back to the factory setting re-enabled the LSI SAS card.

6. Auto Spin-up can be delayed or even stopped given all these drive safety features to prevent a restarted system from overloading the circuit or power supply by jumping all the drives to full speed instantaneously. Regardless, the drive should register. I know there is the right amount of voltage and more than enough current (amperage required to spin up an optical drive motor at full speed is many times over what it takes to run a 7200 RPM platter - and my opticals run fine in my FreeNAS rig).

I checked the Supermicro website, and this exact model of HDD does appear on the list of officially supported hardware. Supermicro confirmed in a support ticket that everything I have done sounds correct and it *should* work. Nothing special should be required in the BIOS settings. These devices should be strictly plug n play. The only other recommendation that had for me was to try backplaning if the drives came back RMAed and still did not function.

One observation that I thought was odd - regardless of whether or not any SATA drives are connected to the SATA Controller or the LSI SAS Controller - in the BIOS the SATA slots always appear empty, and the LSI SAS Controller appears populated. Which seems weird if I have the data cables physically connected to the SATA slots (The white black ones) and not the SAS slots (the blue ones). Not sure if that gives any insight at all? My inclination is that when running the IT version of the firmware, does the SATA controller "assume control" of the SAS ports and adjusts all communication via those ports to be SATA "style"? My own research seems to suggest you can use an SATA drive on an SATA Contoller or SAS Controller, but you cannot use a SAS Drive on anything but an SAS Controller. I'll experiment with some of the firmwares and see. So with all that said, is there anyone here who uses the IT Firmware with SAS drives?

Any ways, when I disable the MOBO SATA I should still see the LSI SAS Controller and vice versa. But in this case, if I disable the SATA I loose both. If I just de-jumper the LSI, the SATA still appears. The whole thing just seems strange as the two controllers should be treated independently (one being the MOBO chipset controller, the other being a PCIe add-on albeit hard wired).

 

[Jailer]

Boy with all the debugging you've done there's not many options left. Is there any chance you could loan one of the drives out and have someone else check if it is detected on a known good working SAS controller?