Highpoint 7102

[joevt]

Actually I'm not sure how 7102 boot.
I post "nvram -p" output here, can you find some clue?

----------------------------------------------------------------
efi-backup-boot-device <array><dict><key>IOMatch</key><dict><key>IOProviderClass</key><string>IOMedia</string><key>IOPropertyMatch</key><dict><key>UUID</key><string>6D1E747C-DC7F-418B-8BB2-B730F377971A</string></dict></dict><key>BLLastBSDName</key><string>disk3s2</string></dict></array>%00

efi-boot-device <array><dict><key>IOMatch</key><dict><key>IOProviderClass</key><string>IOMedia</string><key>IOPropertyMatch</key><dict><key>UUID</key><string>83B132E0-B23E-4857-A02F-83C2B9AF053A</string></dict></dict><key>BLLastBSDName</key><string>disk10s2</string></dict></array>

BootCampHD %02%01%0c%00%d0A%03%0a%00%00%00%00%01%01%06%00%02%1f%03%12%0a%00%00%00%00%00%00%00%7f%ff%04%00

efi-boot-device-data %03%01%08%00%00%01%00%00%04%01*%00%02%00%00%00(@%06%00%00%00%00%00%d8%df%96;%00%00%00%00%e02%b1%83>%b2WH%a0/%83%c2%b9%af%05:%02%02%7f%ff%04%00

efi-backup-boot-device-data %02%01%0c%00%d0A%03%0a%00%00%00%00%01%01%06%00%00%01%01%01%06%00%00%00%01%01%06%00%00%04%01%01%06%00%00%00%03%12%0a%00%01%00%00%00%00%00%04%01*%00%02%00%00%00(@%06%00%00%00%00%00@%ceRt%00%00%00%00|t%1em%7f%dc%8bA%8b%b2%b70%f3w%97%1a%02%02%7f%ff%04%00

I wrote a program to convert those device paths from binary to text.

Your backup boot device is:
PciRoot(0x0)/Pci(0x1,0x0)/Pci(0x0,0x0)/Pci(0x4,0x0)/Pci(0x0,0x0)/Sata(0x1,0x0,0x0)/HD(2,GPT,6D1E747C-DC7F-418B-8BB2-B730F377971A,0x64028,0x7452CE40)

Your boot device is:
Ata(Primary,Slave,0x0)/HD(2,GPT,83B132E0-B23E-4857-A02F-83C2B9AF053A,0x64028,0x3B96DFD8)

The Boot Camp setting points to a disk for BIOS MBR boot:
PciRoot(0x0)/Pci(0x1F,0x2)/Sata(0x0,0x0,0x0)

If you set your Startup Preferences to boot a Boot Camp disk, then the boot device would be set to something like the following (this is from my MacPro3,1):
%01%03%18%00%0b%00%00%00%00%00%e0%ff%00%00%00%00%ff%ff%f7%ff%00%00%00%00%04%06%14%00%eb%85%05+%b8%d8%a9I%8b%8c%e2%1b%01%ae%f2%b7%7f%ff%04%00

which is interpreted as this:
MemoryMapped(0xB,0xFFE00000,0xFFF7FFFF)/FvFile(2B0585EB-D8B8-49A9-8B8C-E21B01AEF2B7)

which means that the Startup Manager will use a program in the firmware with that GUID to start the BIOS compatibility mode which will boot the MBR pointed to by the BootCampHD setting.

But never mind that BootCamp stuff.

That ATA disk seems interesting. Double check that it is the actual value for the RAID by changing the Startup Disk preference to a different disk, check that the efi-boot-device-data has changed, then change the startup disk to the raid again and check that the efi-boot-device-data has been restored to the previous value.
[doublepost=1558297645][/doublepost]

I am not familiar with the terminal command, how to show the full lspci? and how to make sure my correct PCIe device domains to change the 5:* and 6:*?

I use terminal and type code you provide here, it shows command not found.

The ioreg command should work, and will give us more info than lspci.

 

[Mac_User 0101]

Big issue!
SSD7101-A's MacOS driver works on SSD7102, I use "HighPoint RAID Management" to create array, guess what?
Awesome!
View attachment 837551

In Disk Utility, you can find a external drive model name listed as HPT DISK 3_0 Media, then I erase/format it and named as SamsungRaid0, no issue.
View attachment 837552

I test R/W speed of the SamsungRaid0 drive, the performance seems great, maybe the same as SSD7101-A.
View attachment 837553
View attachment 837554


Then I use CCC to clone my MacOS High Sierra 10.13.6 boot drive to this SamsungRaid0 drive and reboot from it, succeed!
View attachment 837555

I try to boot from other drive without install SSD7101-A driver, the SamsungRaid0 drive is not work and disappear in Disk Utility.
And more, I clone a boot drive with no SSD7101-A driver to the SamsungRaid0 drive, it is not bootable.
So I think maybe clean installation of MacOS on SSD7102 array drive and boot from it is not useful, because of no SSD7101-A driver on it.

That's my test report.

So when you were in the HPT NVMe Manager did you notice an option for RAID 5 & 10? Highpoint lists the SSD7102 as only supporting RAID 0 & 1. I'm thinking since it's working with SSD7101-A driver and NVMe Manager that you could possibly have access to the additional RAID levels (they may be greyed out if you don't have 4 identical m.2 installed but you should see it). Did the NVMe Manager report the card correctly as 7102 or 7101-A?
[doublepost=1558332194][/doublepost]

Interesting, I don't understand how the drives are merged as macOS sees just only drive. I'll test this with my SSD7101-A to see if I can do the same.

Have you had a chance to test your 7101-A yet to see if can join the drives with a hardware RAID? I'm dying to find out if this this unique to 7102 or not.

 

[wjesse]

The ioreg command should work, and will give us more info than lspci.

I use the ioreg command and export text as a file.

I think it's two functions (0 and 1) of the same device (16,0). Only three M.2 devices are connected. One is AHCI and the other are NVMe. The bus numbers of the Slot-2 devices seems strange - I would expect that the M.2 devices would use consecutive bus numbers. The device tree may be interesting. I think the following command should give all the information about Slot-2 devices (PXS2 is the name for slot 2 on my MacPro3,1):

ioreg -fiw0 -n PXS2 -rl > ioregslot2.txt

I finally find out my ioregslot2.txt.
[doublepost=1558364168][/doublepost]

So when you were in the HPT NVMe Manager did you notice an option for RAID 5 & 10? Highpoint lists the SSD7102 as only supporting RAID 0 & 1. I'm thinking since it's working with SSD7101-A driver and NVMe Manager that you could possibly have access to the additional RAID levels (they may be greyed out if you don't have 4 identical m.2 installed but you should see it). Did the NVMe Manager report the card correctly as 7102 or 7101-A?

now I'm in the the HighPoint RAID Management, but "Array type" of Create Array function shows nothing.
No devices available to create array, I think I have to delete my raid0 array first.

HighPoint RAID Management report the card as "Model Name: HighPoint NVMe RAID Controller", neither 7102 nor 7101-A.

 

[wjesse]

Do you want to see something funny :
[doublepost=1558285346][/doublepost]

Discontinued?

I think there is a new SSD7103 support more raid type to improve SSD7102.
http://www.highpoint-tech.com/USA_new/CS-product_nvme.htm#bottom

SSD7102
http://www.highpoint-tech.com/USA_new/series-ssd7102-overview.htm

SSD7103
http://www.highpoint-tech.com/USA_new/series-ssd7103-overview.htm

 

[joevt]

I use the ioreg command and export text as a file.



I finally find out my ioregslot2.txt.
[doublepost=1558364168][/doublepost]

now I'm in the the HighPoint RAID Management, but "Array type" of Create Array function shows nothing.
No devices available to create array, I think I have to delete my raid0 array first.

HighPoint RAID Management report the card as "Model Name: HighPoint NVMe RAID Controller", neither 7102 nor 7101-A.

I've reordered and summarized the info:

g1 = 2.5 GT/s (ignore these, as they are internal to the Thunderbolt controller)
g2 = 5.0 GT/s
g3 = 8.0 GT/s

+-o PXS2@0 (g2x16) (PEX 8747 48-Lane, 5-Port PCI Express Gen 3 Switch) (Upstream)

A   +-o pci-bridge@8 (g3x4) (48-Lane, 5-Port PCI Express Gen 3 Switch) (Downstream)
        +-o pci144d,a801@0 (g3x4) (Samsung Electronics Co Ltd NVMe SSD Controller) (AHCI SM951/PM951?)
          +-o AppleAHCI
                      +-o SAMSUNG MZHPV256HDGL-00000 Media
                          +-o EFI System Partition@1
                          +-o Untitled 2@2
                                +-o AppleAPFSContainer
                                  +-o MacOS 10.13.6@1
                                  +-o Preboot@2
                                  +-o Recovery@3
                                  +-o VM@4

B   +-o pci-bridge@10 (g3x4) (48-Lane, 5-Port PCI Express Gen 3 Switch) (Downstream)

E       +-o pci-bridge@0 (g3x4) (PEX 8724 24-Lane, 6-Port PCI Express Gen 3 Switch) (Upstream)

F           +-o pci-bridge@2 (g2x1) (PEX 8724 24-Lane, 6-Port PCI Express Gen 3 Switch) (Downstream)
                +-o pci1b4b,7102@0 (g2x1) (Marvell SATA Controller) (HighPoint 7102?)
                +-o pci1b4b,91a4@0,1 (g2x1) (Marvell 88SE912x IDE Controller)

G           +-o pci-bridge@1 (g3x4) (PEX 8724 24-Lane, 6-Port PCI Express Gen 3 Switch) (Downstream)

H               +-o pci-bridge@0 (g3x4) (JHL7540 Thunderbolt 3 Bridge [Titan Ridge 4C 2018]) (Upstream)
                    +-o pci-bridge@1 (g1x4) (JHL7540 Thunderbolt 3 Bridge [Titan Ridge 4C 2018]) (Downstream)
                    +-o pci-bridge@4 (g1x4) (JHL7540 Thunderbolt 3 Bridge [Titan Ridge 4C 2018]) (Downstream)
                    +-o pci-bridge@0 (g1x4) (JHL7540 Thunderbolt 3 Bridge [Titan Ridge 4C 2018]) (Downstream)
                        +-o pci8086,15eb@0 (g1x4) (JHL7540 Thunderbolt 3 NHI [Titan Ridge 4C 2018])
                            +-o AppleThunderboltNHIType3
                    +-o pci-bridge@2 (g1x4) (JHL7540 Thunderbolt 3 Bridge [Titan Ridge 4C 2018]) (Downstream)
                        +-o pci8086,15ec@0 (g1x4) (JHL7540 Thunderbolt 3 USB Controller [Titan Ridge 4C 2018])
                          +-o AppleUSBXHCITR@01000000
                            +-o AppleUSB30XHCIPort@01300000
I                             +-o USB to PCIE Bridge@01300000
                                      +-o IOUSBMassStorageUASDriver
                                        +-o IOSCSITargetDevice
                                            +-o org_dungeon_driver_IOSATDriver
                                                  +-o JMicron Generic Media
                                                    +-o IOFDiskPartitionScheme
                                                      +-o Untitled 1@1

C   +-o pci-bridge@9 (g3x4) (48-Lane, 5-Port PCI Express Gen 3 Switch) (Downstream)
        +-o pci144d,a808@0 (g3x4) (NVMe SSD Controller SM981/PM981)
          +-o RS7101Controller  <class IORegistryEntry:IOService:RSNVMeController:RS7101Controller>

D   +-o pci-bridge@11 (g3x4) (48-Lane, 5-Port PCI Express Gen 3 Switch) (Downstream)
        +-o pci144d,a808@0 (g3x4) (NVMe SSD Controller SM981/PM981)
          +-o RS7101Controller  <class IORegistryEntry:IOService:RSNVMeController:RS7101Controller>

I used my ParseIORegPCILinkStatus.sh script to interpret the link status values for link width and speed (these might not match the current status).

Connected devices (I won't bother to check if you've already told us what the devices are):
- (A) An AHCI M.2 drive (AHCI SM951/PM951?)
- (C,D) Two SM981/PM981 NVMe drives controlled by the Highpoint RAID driver.
- (H) A GC-TITAN-RIDGE with a USB device (I) connected. The USB device (JMicron?) contains a PCIe device (NVMe?).

Some notes:
The 48 lane bridge (in slot 2, PXS2, of the Mac Pro) only has 5 ports (1 upstream, 4 downstream), which means one of the NVMe slots must be and is controlled by the 24 lane bridge (E). It might be interesting to see how the performance differs between that NVMe slot and the other three. You've chosen to connect the GC-TITAN RIDGE to this slot (H), which might be a good idea, because it's not as fast as your M.2 drives.

The bridges have more lanes than is required by the implementation. Either a bridge with fewer lanes doesn't exist, or the chosen bridge was less expensive (because it doesn't have some unneeded features, such as more ports, etc.). Somehow, a 48 lane bridge with more ports is more expensive than the combination of 48 lane 5 port bridge plus 24 lane bridge or there's another reason for using two bridges.

None of the devices appear to have an option rom ("reg: 48" for devices and "reg: 56" for bridges). Maybe it's hidden somehow, or RAID booting is provided by firmware of the Marvell controller (F) (I don't know how that would work).

Do you actually select a volume on the RAID for boot in the Startup Disk System Preferences panel? See previous post for nvram -p boot setting test instructions.

 

[Mac_User 0101]

Discontinued?

I think there is a new SSD7103 support more raid type to improve SSD7102.
http://www.highpoint-tech.com/USA_new/CS-product_nvme.htm#bottom

SSD7102
http://www.highpoint-tech.com/USA_new/series-ssd7102-overview.htm

SSD7103
http://www.highpoint-tech.com/USA_new/series-ssd7103-overview.htm

Funny thing is, I ordered 7102 last night from Highpoint e-store. I believe I got the last one because it said out of stock afterwards. I contacted them just now and told them to hold the order because I might want 7103. They were very helpful and now I'm speaking with them via email regarding the changes to the card before I take a blind leap of faith. I specifically want to know if it still has (x2) PEX bridge and which ones. I'll report what I find out.

 

[tsialex]

Some notes:
The 48 lane bridge (in slot 2, PXS2, of the Mac Pro) only has 5 ports (1 upstream, 4 downstream), which means one of the NVMe slots must be and is controlled by the 24 lane bridge (E). It might be interesting to see how the performance differs between that NVMe slot and the other three. You've chosen to connect the GC-TITAN RIDGE to this slot (H), which might be a good idea, because it's not as fast as your M.2 drives.
.

It's M1, the top M.2 slot. PEX8714 vias on the back go to the top M.2 slot on the front.

 

[Mac_User 0101]

now I'm in the the HighPoint RAID Management, but "Array type" of Create Array function shows nothing.
No devices available to create array, I think I have to delete my raid0 array first.

I'm totally wondering if 7102 is limited to RAID 0, 1 hardware level or using the 7101-A Manager GUI allows RAID 5, 10 also.

HighPoint RAID Management report the card as "Model Name: HighPoint NVMe RAID Controller", neither 7102 nor 7101-A.

That's a pretty good sign. Makes me believe the driver and NVMe Manger GUI may be universal for 7100 series. They probably don't want us to know that.

EDIT: I have confirmed that the webGUI - NVMe Manger is in fact the same file and universal for entire 7100 series. The drives are also the same with the exception of 7110 which has a larger file size due to SATA/SAS capability. Also, there is no Mac driver for 7102 but the Windows/Linux drivers for that card are the same as 7101-A/7120. The drivers for 7103 are not available for download yet but they have them listed. There appears to be some additional Linux drivers added for that card but no Mac drivers.

 

[Mac_User 0101]

None of the devices appear to have an option rom ("reg: 48" for devices and "reg: 56" for bridges). Maybe it's hidden somehow, or RAID booting is provided by firmware of the Marvell controller (F) (I don't know how that would work).

I believe I found something that pertains to this. I'm not sure how accurate it is or if it answers your question but I'll paste it below:

"Actually, the best way to go through the mentioned limitation is to add a PCIe switch; that’s why PLX8747 chip and custom firmware are used by HighPoint for booting. To be honest, the PCIe switch is expensive, but it does worth its price for allowing configurable interface (located between the drives and the CPU) to work in all cases.

HighPoint has already released a device (the SSD7101A) on the market for allowing 4 M.2 NVMe drives to be connected to the device. Why does it launch SSD7102 now? Of course, this one is even better; the firmware in the PLX chip is changed currently for allowing the boot from a RAID of NVMe drives."
[doublepost=1558405684][/doublepost]I believe the above can probably be done on 7101-A as well (custom firmware or not). As long as you are creating a RAID array with the HPT NVMe Manger and it's done hardware level it "should" be able to boot Mac OS. Hopefully tsialex can confirm for us. (Perhaps this custom firmware is more important for Windows?)

 

[joevt]

I believe I found something that pertains to this. I'm not sure how accurate it is or if it answers your question but I'll paste it below:

"Actually, the best way to go through the mentioned limitation is to add a PCIe switch; that’s why PLX8747 chip and custom firmware are used by HighPoint for booting. To be honest, the PCIe switch is expensive, but it does worth its price for allowing configurable interface (located between the drives and the CPU) to work in all cases.

HighPoint has already released a device (the SSD7101A) on the market for allowing 4 M.2 NVMe drives to be connected to the device. Why does it launch SSD7102 now? Of course, this one is even better; the firmware in the PLX chip is changed currently for allowing the boot from a RAID of NVMe drives."
I believe the above can probably be done on 7101-A as well (custom firmware or not). As long as you are creating a RAID array with the HPT NVMe Manger and it's done hardware level it "should" be able to boot Mac OS. Hopefully tsialex can confirm for us. (Perhaps this custom firmware is more important for Windows?)

I think that's all information we've heard before. It doesn't explain how the firmware works at what the Marvell controller is for in the 7102. I think we have to go looking in the EFI Shell for clues (PCI information, device list, etc.)

 

[wjesse]

Connected devices (I won't bother to check if you've already told us what the devices are):
- (A) An AHCI M.2 drive (AHCI SM951/PM951?)
- (C,D) Two SM981/PM981 NVMe drives controlled by the Highpoint RAID driver.
- (H) A GC-TITAN-RIDGE with a USB device (I) connected. The USB device (JMicron?) contains a PCIe device (NVMe?).

Do you actually select a volume on the RAID for boot in the Startup Disk System Preferences panel? See previous post for nvram -p boot setting test instructions.

- (A) Samsung SM951 AHCI M.2 PCIe SSD 256GB
- (C,D) Samsung 970 Pro NVMe M.2 PCIe SSD 512GB
- (H) M.2 NVMe to USB 3.1 C (Gen 2) Enclosure External PCI-e NVMe SSD Reader Adapter(JMicron chip) with Intel 660p NVMe M.2 PCIe SSD 512GB

Do you mean select a volume on the RAID for boot in the Startup Disk System Preferences panel and run the command "ioreg -fiw0 -n PXS2 -rl > ioregslot2.txt" again?

OK I have done.

 

[startergo]

Somebody tried this:
http://highpoint-tech.com/USA_new/series-ssd6540m-specification.htm

 

[joevt]

Do you mean select a volume on the RAID for boot in the Startup Disk System Preferences panel and run the command "ioreg -fiw0 -n PXS2 -rl > ioregslot2.txt" again?
OK I have done.

I meant the nvram -p command. I think you need to leave the Startup Disk preferences after making a change for the change to be saved to nvram. Then the nvram -p command will show the new value for the efi-boot-device-data.

 

[wjesse]

I meant the nvram -p command. I think you need to leave the Startup Disk preferences after making a change for the change to be saved to nvram. Then the nvram -p command will show the new value for the efi-boot-device-data.

OK I have done.

Before change: nvram -p boot from Mx500.zip
After change: nvram -p boot from Mx500 change to raid0.zip

 

[startergo]

Ok so the SSD7110 oficially supports boot RAID for cMP:
" Thanks for contacting HighPoint.
It does support boot RAID function for latest Mac OS version but have to use on Non-Thunderbolt Mac System or 2012 Mac Tower.
Because Apple Thunderbolt based Mac Systems won’t allow to boot from 3rd paty NVMe Device.
This controller support 3x m.2 and 16x SAS/SATA device."

 

[joevt]

OK I have done.

Before change: nvram -p boot from Mx500.zip
After change: nvram -p boot from Mx500 change to raid0.zip

The Mx500 is disk2s2 and has UUID CDC2CB1C-3F94-49E8-B6FB-4EFEEF7916DC.
The efi device path is:
PciRoot(0x0)/Pci(0x1F,0x2)/Sata(0x2,0x0,0x0)/HD(2,GPT,CDC2CB1C-3F94-49E8-B6FB-4EFEEF7916DC,0x64028,0x7456CE40)

The RAID is as you showed before. It is disk10s2 and has UUID 83B132E0-B23E-4857-A02F-83C2B9AF053A. The efi device path is:
Ata(Primary,Slave,0x0)/HD(2,GPT,83B132E0-B23E-4857-A02F-83C2B9AF053A,0x64028,0x3B96DFD8)

The disk numbers (disk2 and disk10 and whatever) may change between boots (depending on what order the devices are connected or enumerated), but the disk number is not part of the efi device path.

One thing I should have noticed from your original post of ioreg is that the raid volume is not listed under any of the slot 2 devices. This might explain why the efi device path does not appear to be a full path (it starts with Ata instead of PciRoot). That means the "Ata" node of the path might not mean anything. I guess the partition node (type HD, partition 2, GPT partition scheme, UUID, start, length in blocks 511.9 MB) is all that's required for the EFI booter to find the device. First the EFI RAID driver (wherever and whatever that is) creates the EFI disk node then the EFI GPT driver creates the partitions. Then the booter can find the boot partition. It's an HFS or APFS partition which Apple's booter can read to find the boot file.

Here's a command to output the UUIDs of all your partitions:

eval $(diskutil list | sed -nE '/.*(disk[0-9]+s[0-9]+)$/s//printf "\1\\n"; diskutil info \/dev\/\1 | grep UUID;/p')

Give us the results of the following commands (assuming disk10 is still the RAID disk):

diskutil info /dev/disk10
diskutil info /dev/disk10s2

Use the following command to get all the ioreg info then manually search it for the raid device node ("BSD Name" = "disk10" or whatever). Give us the info for that device and all its parent and child nodes.

ioreg -fiw0 -rl > ioregall.txt

I would give instructions for looking at EFI but you would need to install rEFInd and the EFI Shell and I don't know if we want to go down that rabbit hole.

 

[Mac_User 0101]

Discontinued?

I think there is a new SSD7103 support more raid type to improve SSD7102.
http://www.highpoint-tech.com/USA_new/CS-product_nvme.htm#bottom

SSD7102
http://www.highpoint-tech.com/USA_new/series-ssd7102-overview.htm

SSD7103
http://www.highpoint-tech.com/USA_new/series-ssd7103-overview.htm

Highpoint was able to change my order from SSD7102 to SSD7103. I'm all paid for and just awaiting for them to be officially in stock. Based on the info I received directly from Highpoint, they are using 2x PEX 8747 on this card. They also claim improved compatibly on all platforms and NVMe's. Once I receive the card and test it I will start a SSD7103 thread (unless someone beats me to it).

 

[crjackson2134]

Highpoint was able to change my order from SSD7102 to SSD7103. I'm all paid for and just awaiting for them to be officially in stock. Based on the info I received directly from Highpoint, they are using 2x PEX 8747 on this card. They also claim improved compatibly on all platforms and NVMe's. Once I receive the card and test it I will start a SSD7103 thread (unless someone beats me to it).

I’m still trying to understand what possible advantages (over and above the SSD7101A-1) it brings to a cMP user booting only macOS from this card. The hardware I suppose is “New and Improved”, but what will it do for me exactly?

 

[Mac_User 0101]

I’m still trying to understand what possible advantages (over and above the SSD7101A-1) it brings to a cMP user booting only macOS from this card. The hardware I suppose is “New and Improved”, but what will it do for me exactly?

Honestly, not much at all. Especially if you don't plan on building a bootable hardware RAID array either. I still don't know for sure if the 7101-A can do the same. May be possible to use the HPT driver and webGUI to create a hardware RAID array and also boot Mac OS like the SSD7102 "unofficially" does. No one has confirmed or unconfirmed that for me yet. I pretty much decided to go with the new SSD7103 because I'm confident it will work similar for cMP as 7101-A/7102 does and I'm kind of excited to test the new card.

Apparently with the SSD7103, it's just as compatible with AMD processors now as it is with Intel. That might be good for me if I ever build a Ryzen Hackintosh. Should have a quieter variable speed fan. Highpoint also claims that they went through great lengths testing just about every NVMe currently available for compatibility and stability. It should be a solid card but nothing to run out and switch your 7101-A for.

 

[edgerider]

I’m still trying to understand what possible advantages (over and above the SSD7101A-1) it brings to a cMP user booting only macOS from this card. The hardware I suppose is “New and Improved”, but what will it do for me exactly?

this I can explain...
i use true hardware raid since 2008, and would not consider any workstation without it.

hardware raid dont only let you built large array or aggregate disk...

the most important feature is to offload disk read and write to the raid controler.

your machine becomes way more snappier, and if you have a true and powerful raid card with a decent ram cache, it is then even faster.

my 24bay raid 50 sas array can do 2gb/s read write all day long 24/7, and there is no noticeable difference in latency compared to my anfeltec squid with 4sm951.

the other good point is that with a hardware raid controller the raid set is on the drives, not on the controler.

if your raid card go kaboom, you just replace it and put back the drives on it and you are golden.

last but not least hardware raid controller are able to perform disk to disk clone faster than any sytem, and you can do migration of raid level, or migration of drive in size.

therefore i would totally dig a 16 m2 slot card with a true hardware raid controller
to replace my 24 drive array, because there is one thing server grade raid cant do : being ****ing silent and consume less power ...

when in use my xserve with the two 24bay raid array are doing around 2400w on 220v so if i could have the same thing getting rid of the 3.5” sas platter, I will totally love this!

but trust me the moment you try a true hardware raid you cant do without it.

it saved my ass a couple of time and allow me to deal with very large video project that most of other cant just handle at once and as fast.

 

[Mac_User 0101]

this I can explain...
i use true hardware raid since 2008, and would not consider any workstation without it.

hardware raid dont only let you built large array or aggregate disk...

the most important feature is to offload disk read and write to the raid controler.

your machine becomes way more snappier, and if you have a true and powerful raid card with a decent ram cache, it is then even faster.

my 24bay raid 50 sas array can do 2gb/s read write all day long 24/7, and there is no noticeable difference in latency compared to my anfeltec squid with 4sm951.

the other good point is that with a hardware raid controller the raid set is on the drives, not on the controler.

if your raid card go kaboom, you just replace it and put back the drives on it and you are golden.

last but not least hardware raid controller are able to perform disk to disk clone faster than any sytem, and you can do migration of raid level, or migration of drive in size.

therefore i would totally dig a 16 m2 slot card with a true hardware raid controller
to replace my 24 drive array, because there is one thing server grade raid cant do : being ****ing silent and consume less power ...

when in use my xserve with the two 24bay raid array are doing around 2400w on 220v so if i could have the same thing getting rid of the 3.5” sas platter, I will totally love this!

but trust me the moment you try a true hardware raid you cant do without it.

it saved my ass a couple of time and allow me to deal with very large video project that most of other cant just handle at once and as fast.

SSD7101-A can build a hardware RAID too using the Highpoint driver and webGUI (NVMe Manager). It allows for RAID levels 0, 1, 5, 1/0 but the array is supposedly not bootable under any OS. It is currently the only Highpoint carrier that can do cross RAID feature allowing multiple 7101-A's to be combined into a single array. SSD7102 and soon to be 7103 have "custom firmware" on PEX switch that allows booting. SSD7120 is similar but for U.2. The RAID bootable SSD7102 & AFIK 7103 don't officially support Mac OS. I found out thanks to wjesse's testing that we can use the Mac driver & webGUI for 7101-A on the 7102 to build the array which is bootable on Mac OS. Most people don't use their 7101-A like that and opt for Disk Utily RAID between x3 m.2 while booting to x1 m.2. It may be possible to boot the hardware array on 7101-A too but no one has confirmed that to me to yet. I also have a feeling that cross RAID may be possible with the other cards but that is also unconfirmed. I won't know that until someone actually tries it.

 

[startergo]

SSD7101-A can build a hardware RAID too using the Highpoint driver and webGUI (NVMe Manager). It allows for RAID levels 0, 1, 5, 1/0 but the array is supposedly not bootable under any OS. It is currently the only Highpoint carrier that can do cross RAID feature allowing multiple 7101-A's to be combined into a single array. SSD7102 and soon to be 7103 have "custom firmware" on PEX switch that allows booting. SSD7120 is similar but for U.2. The RAID bootable SSD7102 & AFIK 7103 don't officially support Mac OS. I found out thanks to wjesse's testing that we can use the Mac driver & webGUI for 7101-A on the 7102 to build the array which is bootable on Mac OS. Most people don't use their 7101-A like that and opt for Disk Utily RAID between x3 m.2 while booting to x1 m.2. It may be possible to boot the hardware array on 7101-A too but no one has confirmed that to me to yet. I also have a feeling that cross RAID may be possible with the other cards but that is also unconfirmed. I won't know that until someone actually tries it.

I wonder is there bigger benefit of 7110 over 7103, taking into account the fact that the 7110 is officially supported bootable in hardware raid, but also is almost double in price. Plus 7110 supports 3 M2 NVMe + 16 SAS/SATA devices

 

[wjesse]

Here's a command to output the UUIDs of all your partitions:

eval $(diskutil list | sed -nE '/.*(disk[0-9]+s[0-9]+)$/s//printf "\1\\n"; diskutil info \/dev\/\1 | grep UUID;/p')

Give us the results of the following commands (assuming disk10 is still the RAID disk):

diskutil info /dev/disk10
diskutil info /dev/disk10s2

Use the following command to get all the ioreg info then manually search it for the raid device node ("BSD Name" = "disk10" or whatever). Give us the info for that device and all its parent and child nodes.

ioreg -fiw0 -rl > ioregall.txt

The first command shows:
-------------------------------------------------------------------------

-bash: ??diskutil: command not found
sed: illegal option -- /
usage: sed script [-Ealn] [-i extension] [file ...]
sed [-Ealn] [-i extension] [-e script] ... [-f script_file] ... [file ...]

-------------------------------------------------------------------------

I upload disk10, disk10s2 and ioregall.txt files, but I'm not sure why the ioregall.txt shows nothing.
[doublepost=1558537225][/doublepost]

I wonder is there bigger benefit of 7110 over 7103, taking into account the fact that the 7110 is officially supported bootable in hardware raid, but also is almost double in price. Plus 7110 supports 3 M2 NVMe + 16 SAS/SATA devices

Before I bought 7102, I emailed to Highpoint and asked them if 7102 support MacOS.

this is their answer.
----------------------------------------------------------------------------------------
We suggest you to buy SSD7110. Here is our website information http://highpoint-tech.com/USA_new/series-ssd7110-download.htm

It support Mac OS X from 10.12 and later.

HighPoint NVMe SSD doesn’t support AHCI protocol devices.
-----------------------------------------------------------------------------------------

So I think they prefer Mac user to buy SSD7110.

 

[joevt]

The first command shows:
-bash: ??diskutil: command not found

It should have worked. Did you type it instead of copy and paste it?
Give me the output of "diskutil list > diskutillist.txt". Maybe there's something weird in that which made the original command fail.

I upload disk10, disk10s2 and ioregall.txt files, but I'm not sure why the ioregall.txt shows nothing.

You should pipe the output of the commands to text file with something like "> disk10.txt", etc. Text files should have .txt file extension (doesn't matter for the files you already uploaded since I can view them anyway). Not sure why you uploaded a 0 byte file. I made a mistake in the ioreg command. It should be "ioreg -lfiw0 >ioregall.txt".

There's nothing unexpected in the disk10 and disk10s2 diskutil info results. The device is identified as using PCI-Express protocol. The "Partition UUID" is used in EFI device paths - I think it comes from GPT partition table. The "Volume UUID" is probably from the HFS+ volume info.

About my previous comment about the raid volume not appearing under the slot 2 devices: I should have realized that it wouldn't be listed there because a RAID volume consists of multiple devices and therefore will not be listed under a specific device (only the parts of the RAID will be listed under a device) - I'm guessing the RAID volume will be listed under IOResources.

I'm also guessing that the bless command (or whatever the Startup Disk preferences panel uses) is defaulting the device node path to "Ata(Primary,Slave,0x0)" when the volume is a RAID (not attached to a single device). I think this is different from an Apple software RAID which would boot a small partition that would load the RAID.

 

[Mac_User 0101]

I wonder is there bigger benefit of 7110 over 7103, taking into account the fact that the 7110 is officially supported bootable in hardware raid, but also is almost double in price. Plus 7110 supports 3 M2 NVMe + 16 SAS/SATA devices

SSD7110 is a great if you need the SAS/SATA in addition to 3x m.2 slots all from a single slot. I was interested in this card and was contemplating buying it for m.2 carrier and external RAID array. Highpoint pretty much told me it was a 7102 with 16 internal SAS/SATA via 4x SFF-8643 ports. Classic Mac Pro 3,1 could use this card to upgrade internal bays to SATA 3 by going mini SAS to molex plugged into the backplane. 4,1 and 5,1 could technically install drives backwards in internal slots and go mini SAS to 4x SATA breakout cable and take the power from the internal SATA ports for each drive. I've never actually tried that but I'm pretty confident it world work.

There is a disclaimer on the Highpoint website for the SSD7110 under RAID support that states "for Mac support, please contact sales." Highpoint explained to me that it's not that they don't want to support Mac with the bootable RAID cards but rather that the classic Mac Pro's are the only Intel Macs with native PCIe slots. Modern T2 equipped Mac's who would be using such a card in a Thunderbolt chassis have external booting disabled by default. Even if you enable it you still can't boot an external HFS+ volume if encryption is enabled. I didn't know this until they explained it because I've never really used a Mac with Thunderbolt. It makes sense to me that universal Mac support could get complicated with all these variables. Please correct me if I'm wrong about any of that.