Wody

The document about them is here, it may just be the standard model. Other potential issues are a wrong sector size, so not 512 or 4096 but one of the others, and the drives could be broken too, although usually there would be some sense codes if so.

I think your drives are WUH722020BL5204 ? If so, they seem to have SED security enabled, you'll need to remove the protection before they are usable. For seagate sas drives, they detect they are attached to a new system and erase themself automatically, until they are done they are not available, this is probably something similar where you either have to tell them to erase themself, or if they are doing it, wait until they are done.

They need more power and generate more heat than the onboard controller, which can be a concern if your case isn't big enough or doesn't have enough airflow.

Generally mixing memory is not recommended. So you can also try and see what happens when you run only micron, and if they give errors, try hynix. The trick is usually to find out the most basic configuration where things run, and then start adding to it, and when it fails, it's usually the last thing you added, whether it's broken, not compatible etc. Since the address is the same all the time, you can also try switching modules, then if the error follows the module, it's something about the module, if the address stays the same, it's the processor, or something between, maybe some dirt or dust so things work most of the time, but not always. It does say 'no action required', so you can ignore it, but it just isn't like how it should be.

The 9400 has its page here and it does have enabler cables, which pinouts are in the manual. I'm not sure how and if they are different on the drive-side. The 9300 (doesn't support NVMe) and the 9400 (and 9500) have a cable guide document here. I can't find a H300, but all the H3xx I did find have chips like the 9300-series or older. SFF-8654 comes in 8i and 4i variants, the 9400 seems to only have 4i cables, the 9500 8i. There's also others than broadcom who make sff-8643 to sff-8654 8i cables, in both SAS/SATA nad NVMe variants which are different, from my understanding though, they don't work as enablers, so you'd need an enabler cable, then some adapter, and then a cable to the backplane, if it uses 8i. Cables are a big mess. If only we had U.3 backplanes already, then all those issues would go away. Apparently the PM1633 has firmware issues, like the 32k bug, so I'd check firmware. Also they support different sector-sizes, so you'd want to make sure they are formatted as 512 per sector, not 520 or something like that. They may also have encryption, which shouldn't be an issue, but could affect performance. My server is more traditional with ST10000NM0096 drives and an M.2 for cache, currently on a 9500, before I ran a 9400 with a SATA SSD as cache. The backup-server has a 9600 with about a dozen drives, with HGST Ultrastar 8TB SAS being the biggest, the rest are old sata-drives. Edit: Also found this knowledge base article about cables which seems more recent and has some tips.

There's usually a slight difference between specs and reality, as long as you haven't enabled overclocking and the bus-speed is stable it should be fine.

IPMI typically only can use onboard ports, and normally it has a dedicated port, which means you would run two cables. If the port is shared and you use two cables, then there is a chance some transfers will go over the slow connection, and some go over the fast one. You can set eth0 to the 10g one in unraid, which means it should use that as a default. As for DHCP, you shouldn't have to do or set anything for things to work (assuming there is a DHCP server on the network) since each port gets its own address.

Since it is a current model, you can find the 9500-16i here, in the documentation section you can find the manual, which describes what cable you need (search for U.2 and it will go right to the cable section). It also has a Storage Adapter Drive Compatibility Report but since that is from 2019, it may or may not help. Unraid generally doesn't care what drives you want to use, but the controller you use might. Backblaze publishes drive reports which may help with reliability, but some good advice is to use different brands, different types, buy them at different times from different stores. Basically, if you buy all the same drives at the same time from the same store, they are likely to be shipped together and made at the same time too, so if one fails, it's pretty much guaranteed the others will fail around the same time too. Of course you should have backups, but that doesn't help much if all your drives keep failing.

RAID means Redundant Array of Inexpensive Disks. For a normal HBA or drive controller, it just passes through drives, but with a RAID controller, they control arrays instead. That usually means they won't pass through random drives, because they are not part of an array, so you have to make them part of an array. If you'd have an Adaptec RAID card, those do pass through drives not part of an Array, but LSI/Broadcom does not (although that may change in the future). Since you have an HP card, I don't know how that handles things. So, you'd have to make arrays of single drives, and set them to RAID 0 which means no data security to pass through the drives as clean as possible (RAID has different ways to keep data safe). So you'd have to make an array, add a drive to it, set the array to RAID 0, store and close the array, make another one, add the next drive and so on. The problem with RAID is that it has to remember information about the array, and it writes that to drives part of the array, which could interfere with parity checks if this data gets changed, but you'll find that out soon enough.

Running a memory-test would show if it was really broken or not. But as far as I know, it can't be a software error, because it's the module itself reporting errors. If it was normal memory, you would get crashes, or corrupted data, or weird errors. As for why would it just break, memory is basically a whole bunch of switches that are on or off, so eventually they break down, plus things like heat can be an issue, and there can be electrical issues. With ECC memory, it verifies that these bits of memory are working correctly, and if not, they throw these errors. It could also be compatibility issue, not all modules are fit for all motherboards, usually it means the board won't boot at all, but all kinds of stuff can happen (for example, I have a broken module where one board says no module is inserted, on another board it boots fine, but run a memtest and it hangs the computer after 2 seconds). So yes, you should be worried and run a memory-test, if that throws error, the memory has to be replaced. You could also run the test on another computer after that, to see if it fails there too, if it does, it's the modules, otherwise it could be something else. Since the memory-controller is on the CPU, it can also be the board that is broken, or the CPU, and it can also be that the torque isn't correct so things aren't making contact properly. But ECC errors usually are the modules.

Those errors say something went wrong with your memory, but because of ECC they were corrected. But this means either your memory is broken, or the CPU (which has the memory controller on it) is. If you had more weird errors, it could be something about the power supply. Memtest86+ doesn't handle ECC I think, but you can try that, the commercial memtest86 does support ECC errors and detects them as well, you could use that to detect them, then switch modules to see if the error follows the module, or stays at the same position. If it's the module(s) replace them, if it stays, your CPU is broken.

B7 means Configuration Reset (reset of NVRAM settings), so try resetting the CMOS. Removing the battery isn't enough, you have to also remove all power and then short the pads on the motherboard used for that purpose, so follow the procedure in the motherboard manual. If that doesn't work, replace the battery, if that doesn't work, the board is probably broken.

If SUM is the Smart Update Manager, that can only be used for firmware, but not to create or destroy RAID arrays, and being old unsupported hardware, I doubt SUM will even do that. The quickspecs for the card mention for that you can use ACU, Array Configuration Utility or the Option ROM Configuration for Arrays (ORCA) which you can reach by pressing F8 during boot. (The product page has lots more documents.) The product page also leads to the latest firmware and installations for that, but I'm having difficulty finding ACU in the list of stuff, so ORCA may be your best bet.

The 9400 does need EFi shell for cross-flashing, otherwise it doesn't. For the 430, you'll need Windows and the firmware .EXE which asks to install or extract, if install doesn't work (because it may verify the computer, not just the card) you can extract and do a manual install. for the 9400, you'd get storcli64, the firmware, and bios-files if required, and in windows, linux, or efishell do storcli64 /c0 file=filename for EFI, storcli64 /c0 efibios file=filename for the efi-interface and storcli64 /c0 bios=filename for the legacy bios. After that you can just run efibios64 /c0 restart, or restart the computer. The hardest part is typing all the actual filenames, so I didn't this time. Usually I also rename them to something simpler, but with tab-completion it doesn't matter much. edit: Interestingly, the 9500 used to be much simpler (just one firmware, and one bios), but they also added the PSOC package, which I think is used to boot the card, and they added mixed mode and sas/sata only with an update to release 30. The 9600 is also much simpler, that uses storcli2 and only one file with everything.

Tri-mode means SAS, SATA and NVMe, as opposed to traditional HBA's that 'only' do SAS and SATA. Flashing the non-NVME firmware probably won't make a difference in performance, but they did fix a lot of bugs. A lot of it will be for the entire product line, and HBAs are pretty simple so they won't have a lot of issues, but generally, you want to use the latest firmware to avoid any known issues. The 9305 and newer no longer have IR firmware, and the 9400 and newer don't have the option to switch,so there are only HBA's and RAID-controllers. You probably will have to flash it, most won't come with the P24 (version 24) firmware, almost nobody flashes stuff. But, flashing is easy, because you can't flash a wrong version, or have to do weird stuff with jumpers or DOS or EFI-shells.

About JBOD/IT mode, it's basically the same thing. In IR mode, the controller acts as a raid-controller which doesn't pass through single disks, where in IT mode, it acts as just a pass-through for just a bunch of disks. (JBOD). 1. The sas 2008 supports SAS2 and Sata 3 speed (6gpbs) so for many drives it may not make a difference. But the 9400-16i has the sas 3416 chip which also supports SAS3. The 2008 has PCIe2.0 and the 9400 has PCIe3.1 so if that was a bottleneck, it will make a difference. Also, the 9400--16i processor runs faster, so even if the transfer-speeds are the same, there should be a difference. It sure feels a lot faster and more responsive to me. 2. For older cards, you'd have to flash them to get into IT mode, but also to fix bugs. For the 9400, it's mostly to fix bugs, since the 9400 and newer are only HBA's so they don't support RAID at all. It does come with two firmwares to choose from, the Mixed one support SAS, SATA and NVMe (with special cables for NVMe), while the SAS/SATA only firmware doesn't support NVMe but has bigger buffers and more queues. 3a. The NVMe firmware won't be slower, but with bigger buffers and queues it can handle more commands at once, and sort things better, so you'd only notice a difference if you use the card at max capacity which with unraid is unlikely to happen. 3b. NVMe is the protocol used to talk to flash drives, it's mostly used for PCIe drives (M.2 is an interface that combines different physical protocols like PCIe and SATA and some others). So if you buy a SATA SSD, it talks SATA, if you buy an M.2 or PCIe drive, it talks NVMe. For a SATA or SAS SSD you don't need an adapter, for the drives that talk NVMe you'd need a special cable. Usually those drives come with an U.2 connection, which looks like a SAS connector but isn't, and a SAS connector looks like a combined power+data SATA connector with the bit in the middle filled in. 4. The latest firmware for SAS/SATA is the Lenovo firmware from here which is version 24.00.07.00 from 2024. The latest firmware Broadcom released is here which is version 24.00.00.00 from 2022.

set/p is not a typo, the SET command is used to set environment variables that can be used by other programs, such as the computer-name for example, or the path where to look for programs. Specifically that instruction (set/p) says to display a prompt "allow uefi" where you would say Y or N, and store the result in the variable named boot. Then the next instruction would check if the boot variable contains 'N' and if so, it would disable the EFI folder by renaming it EFI-, or if it was Y, enable UEFI booting by naming it EFI.

Nope. First description it says PCI-E Ultra-port, and later it mentions the SFF-8654 is for the backplane, and the SFF-8087 is for the controller. I'm not sure if there are any PCIe controllers with SFF-8087, or if it works with SATA/SAS, but they won't work the other way around.

In the old days, you'd get a nice book with the motherboard were they had things like 'does not work with this device, or set this for that device' but these days they don't seem to test much anymore. So since it's a pretty new board, I'd contact MSI, they also have forums, Being a gaming-board, they probably never tested server-stuff for compatibility, so you need to get them to figure out what the issue is.

RAID controllers are normally not recommended, but the adaptec ones should be okay because they can pass-through non-RAID drives. The Adaptec Smartraid 3154-16I got new firmware two months ago, so make sure that is installed. If it still doesn't work, try contacting Adaptec, to see if it is a known problem. Unfortunately the site doesn't seem to work properly, so I can't download the firmware or release notes to see if it is something they fixed.

Mixing SATA and SAS is no problem. The controller negotiates the speed for each device separately, so they all get full speed. In case a drive fails, it could matter since SATA tends to block things. I've not had any broken drives fail while they were powered on though, so not sure what would happen. But for throughput it doesn't matter. The carriers are included with the motherboard, you can verify that with the parts list on the page you listed, although there are three types and only two are included. It says on the processor which one you need (E1B) which is the MCC carrier. The case I'm using uses SFF-8643, so I have SFF-8654 to SFF-8643 cables, and I solved it by getting SFF-8643 to SFF-8087 adapters, and short 8087-cables. Keep in mind a SATA connection can be 1 meter long at most, so if you get a 1 meter cable, then an adapter and then another cable, it may not work.

You don't need a jumper, it is only for certain models of the card. Not sure if your Dell supports UEFI, but you can also use a Freedos USB that you can make with Rufus. You'll need the BIOS and sas3flash from here, and the firmware from here (the firmware from Broadocom has a bug with SATA drives). Assuming you're using an EFI-shell, you use sas3flash, if you're using DOS, it will be sas3flsh. You already got a copy of the SAS-address in this post, but make sure to write it down somewhere in case you need to put it back. The following commands should not erase it though. First erase the flash: sas3flash -c 0 -o -e 6 Then to program the firmware: sas3flash -c 0 -o -f filename -b mptsas3.rom -b mpt3x64.rom (assuming you want the legacy and UEFI bios files as well, otherwise you'd just do the -f filename part.) Now do: sas3flash -list to verify the flash worked, and if it forgot the address sas3flash -o -sasadd 51866XXXXXXXXXXXX (replace 51866XXX with the actual address of course)

1. No. If you'd have a RAID controller there would be issues, because they write some data on the drives that may not be compatible with another one, but a HBA does not have that issue. 2. the 9300 is really two 9300-8i's in disguise, using twice the power. The 9305 has 93 in it, but is a completely new design. The 9400 supports NVMe with special cables, and uses less power, the 9500 has new connectors and uses even less. 3. The 9300 absolutely requires airflow, the newer modules get far less hot, but still need some, it'll be the hottest part in the computer. 4. & 5. Dell has had the habit for decades now to make their computers non-compatible with parts not bought from them. So if you buy a non-Dell model, you should not need any modifications. 6. After verifying the smaller parity drive is no longer needed, stop the array, assign it to a data slot, start the array, and let unraid clear and format it. 7. Depends on the size of the cache-drive. I went from a big SATA one to a smaller NVMe so where it says replace the cache drive, I destroyed the cache pool, started the array, stopped it, made a new cache pool and assigned the NVMe, and then moved everything back. If the NVMe is the same size or bigger, you should be able to just assign it.

The 430-8i you linked to won't work, since that isn't PCIe form factor, but the PCIe version is fine. They use SFF-8643 cables. The Norco uses SFF-8087, so yes you need SFF-8643 to SFF-8087 cables. Make sure to get SAS/SATA cables, there are also NVMe versions that won't work since they are wired differently so if you connect them the card won't detect anything. Unfortunately, there are no SFF-8654 to SFF-8087 that are SAS/SATA, to the best of my knowledge. I bought several a year ago that claimed to be, and they were NVMe. Hopefully that changed though, but if not, there are adapter-cards, from Chenyang which are supposed to take a slot, but you can just remove the backplate and put them almost anywhere. You also may consider getting the 9400 (430) 16i instead, since the 8i uses 10 watt, which means 20 watt for 16, while a 16i only uses 12. Also don't forget the powersupply needs at least 6 molex-connectors. I got a full modular powersupply, that came with only 3, so I ended up buying an additional cable directly from the manufacturer.

There is a 'dense' version of it which obviously won't work, but for the PCIe version it should be as easy as making an USB with storcli and firmware which you can get here, and bios if required, booting into EFI-shell, flash the firmware with fs0: storcli64 /c0 download file=HBA_9400-8i_Mixed_Profile.bin once done, restart the computer, and it should identify as a 9400, if you need the BIOS you can flash those with replacing it with 'download efibios file=' and 'download biosfile=' and of course the respectable files, mpt35sas_x64.bin for UEFI and mpt35sas_legacy.bin for the BIOS. For the firmware, there are options nosigchk and noverchk but those shouldn't be required. As for non-compatible, there may be some firmware check where it refuses to load if it isn't on a non-Thinksystem but I doubt it, seeing how people don't report anything. I think it's more a question of not wanting to support it. As for IT mode, the 9300 has IR and IT mode, but the 9305 and newer models only support IT mode. There are different firmwares for newer models though, the 9400 and 9500 have firmware for SAS/SATA only which has bigger buffers and queues since it doesn't need the NVMe stuff, and mixed mode firmware for SAS/SATA/NVMe. Edit: So I wanted to know if it worked or not, thought about getting a Lenovo model, then I figured, if it works one way, it should work the other too. So I fpund the lenovo firmware and flashed my regular 9400-16i to a 430-16i and that worked. Some notes: If you try to flash the card under linux, it will recognize the firmware (storcli64 show file=firmware.bin) but when you try to flash it, it will say it's invalid, so you really have to use EFI shell. The last official firmware is 24.00.00.00 from 2022 but Intel has a newer firmware 24.00.04.00 from 2023 which you can get here. It includes both SAS/SATA only and mixed SAS/SATA/NVME firmware. If you're reading this after 29 Sept. 2031, sorry, they deleted it. RSP3QD160J is their 9400-16i model, RSP3GD016J is their 9400-16e model. The other one is a mezzanine. Lenovo firmware does something funky in their installer, but the firmware seems standard, and you can get that here. If it asks you for a serial, cancel and press the button again. The firmware is version 24.00.07.00 from 2024, so even newer, but includes only the SAS/SATA versions. Big differences between the Lenovo and standard firmware are bugfixes, a different PCI-id and buffers. The standard firmware does about 6600 Parallel commands, the Lenovo version about 9000. So, if you need NVMe use the standard or Intel version, if you don't, you can use any of them. As long as you don't hold me responsible if something goes wrong, of course. Final note: Broadcom changed their site, to separate legacy products, so now you can go to Support and services, then Documents, Downloads and Support, then Support Documents and Downloads, select Legacy products for Product group, then Legacy Host Bus Adapters (or Legacy Raid controllers if you have one) for Product Family, and then start typing the model number under Product name and it should come up.