Problems with new disk installation

Hardware
1

Today I put in 2 new pieces of hardware

  • StarTech 8P6G-PCIE-SATA-CARD (ASMEDIA ASM1062 controller)
  • Silverstone FS204 4 bay SATA cage

My computer sees the card while booting… the Smart check shows it on the screen and lists the 2 disks I put in the cage.

In the UEFI boot menu the 2 disks in the cage are listed as boot options

But,
When I boot Linux, it sees the card, but not the disks

$lspci
.....
07:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 02)
08:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 02)
09:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 02)
0a:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 02)
0c:00.0 USB controller: ASMedia Technology Inc. ASM1042 SuperSpeed USB Host Controller
0d:00.0 USB controller: ASMedia Technology Inc. ASM1042 SuperSpeed USB Host Controller
0e:00.0 USB controller: ASMedia Technology Inc. ASM1042 SuperSpeed USB Host Controller
0f:00.0 Network controller: Qualcomm Atheros AR9285 Wireless Network Adapter (PCI-Express) (rev 01)
10:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 01)

but
lsblk does not see the 2 disks.
The LED’s are on , on both disks. They have power.

So I conclude… the hardware is done right, but there is something missing from linux.
I looked up the cards here

It says that it needs the ahci module

lsmod | grep ahc
acard_ahci             12288  0
ahci                   49152  6
libahci                61440  2 ahci,acard_ahci
libata                475136  4 libahci,ahci,acard_ahci,ata_generic

I HAVE the ahci module loaded.

So what is missing?
There is heaps in /lib/modules/6.5.0-1mx-ahs-amd64/kernel/drivers… it is hard to guess what is needed.

The card specs are here

It says Linux supported.

3 Likes
2

What does lspci -k say?
It should also print some info about driver used, if any…

Did you try to look at dmesg?
Maybe it gives some hint…

Edit:

Without really blaming, just to exlude the SATA cage as culprit:
can you try to hook up the disks directly to the controller?

2 Likes
3 
$ lspci -k

07:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 02)
	Subsystem: ASMedia Technology Inc. ASM1062 Serial ATA Controller
	Kernel driver in use: ahci
	Kernel modules: ahci
08:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 02)
	Subsystem: ASMedia Technology Inc. ASM1062 Serial ATA Controller
	Kernel driver in use: ahci
	Kernel modules: ahci
09:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 02)
	Subsystem: ASMedia Technology Inc. ASM1062 Serial ATA Controller
	Kernel driver in use: ahci
	Kernel modules: ahci
0a:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 02)
	Subsystem: ASMedia Technology Inc. ASM1062 Serial ATA Controller
	Kernel driver in use: ahci
	Kernel modules: ahci

......
10:00.0 SATA controller: ASMedia Technology Inc. ASM1062 Serial ATA Controller (rev 01)
	Subsystem: ASUSTeK Computer Inc. ASM1062 Serial ATA Controller
	Kernel driver in use: ahci
	Kernel modules: ahci

I think (rev 01) (last item) is another controller… I have 3 one onboard, and one on PCI, and the new one in a PCIE slot.
The driver is ahci. It is loaded.

Can do .
I doubt if cage is problem, the BIOS sees the disks and their brand names appear in the boot menu.

dmesg

[   12.287492] ata11: found unknown device (class 0)
[   12.287507] ata11: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[   12.287514] ata11: link online but 1 devices misclassified, retrying
[   12.287517] ata11: reset failed (errno=-11), retrying in 10 secs
[   22.527493] ata11: found unknown device (class 0)
[   22.527505] ata11: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[   22.527510] ata11: link online but 1 devices misclassified, retrying
[   22.527513] ata11: reset failed (errno=-11), retrying in 35 secs
[   58.415493] ata11: limiting SATA link speed to 3.0 Gbps
[   58.879492] ata11: found unknown device (class 0)
[   58.879503] ata11: SATA link up 3.0 Gbps (SStatus 123 SControl 320)
[   58.879508] ata11: link online but 1 devices misclassified, device detection might fail
[   59.188698] ata12: SATA link down (SStatus 0 SControl 300)
[   59.492697] ata14: SATA link down (SStatus 0 SControl 300)
[   59.796721] ata16: SATA link down (SStatus 0 SControl 300)
[   59.796989] scsi 16:0:0:0: Direct-Access     ATA      ST4000DM004-2U91 0001 PQ: 0 ANSI: 5
[   59.798207] scsi 23:0:0:0: Processor         Marvell  91xx Config      1.01 PQ: 0 ANSI: 5

I dont think it is relevant… I think that is the Marvell sata port
Other than that there is no mention in dmesg

I am going to try booting a gparted CD. gparted within Linux does not see the disks.

Thanks
Neville

1 Like
4

I’m not so doubtful. The cage adds extra connections and extra cable length possibly with a different surge-impedance which can cause reflections and can severly degrade signal quality, thus can create unreliable data transfer between the controller and the disk, especially at high speed.
I’d try to connect the disk directly, just to see…
BIOS maybe uses the contoller with the minimum SATA speed, whereas a booting OS tries to use it at higher speed.

2 Likes
5

@kovacslt ,
Well, gparted, booted from a CD can see the new disks.
I formatted one of them… put an EFI partition and some other partitions.
Then booted back into MX. It still cant see the disk, even after I have formatted it. gparted running within MX can not see the new disks.
so
gparted on the CD has some drivers or whatever , that MX does not have?

and.
This proves that the cage and the controller card are OK. The problem is software… neither MX, nor Void, nor Devuan can see these new disks, but a gparted CD can see them.

I am going back to the gparted CD to see if I can work out what it is using to drive these disks.

4 Likes
6

Great, you just need a different kernel then.

4 Likes
7

I will try a fresh install… it may drag in the correct drivers if the new disks are present when I install.

I could not see anything useful in the CD version of gparted… it is quite old… so this is not a new hardware support issue.

5 Likes
8

Ubuntu to the rescue! :slight_smile:

What is the gparted CD running? Something using systemd?

Just giving you a hard time.

9

Gparted CD is Debian… so yes systemd. Why would that matter? Do you think there is a daemon missing? Hard enough.thanks.
Something as primitive as the BIOS can see the disks and lists them in the boot menu,
it cant require anything sophisticated.

10

I have one cable that came with the card. It reaches the cage easily, there are no extensions. Yes the cage does add another set of contacts.
I can still try it.

11

I meant the cage itself. There must be some wires inside from the outer connector to the inner connector for the drive to be installed.

What I described in my previous post is now just a theory which should be disproved :wink:

This part:

[   22.527505] ata11: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[   22.527510] ata11: link online but 1 devices misclassified, retrying
[   22.527513] ata11: reset failed (errno=-11), retrying in 35 secs
[   58.415493] ata11: limiting SATA link speed to 3.0 Gbps
[   58.879492] ata11: found unknown device (class 0)
[   58.879503] ata11: SATA link up 3.0 Gbps (SStatus 123 SControl 320)
[   58.879508] ata11: link online but 1 devices misclassified, device detection might fail

in my reading tells about that there’s a problem on the link with high speed, reduced to lower speed, still not working: link online but 1 devices misclassified, device detection might fail

This is exactly what you are facing with: device detection failed.

So I’d try to connect at least one disc directly, which does not work in the cage.
If it still does not work, then my theory was failed.
If it does…

Also it would be interesting to know the link speed in your the gparted CD, where the discs work in the cage.
Some hint about how to know it: https://www.cyberciti.biz/faq/linux-command-to-find-sata-harddisk-link-speed/

If the speed there is set to SATA1 (1.5Gbps), I’m most probably right with my theory. Otherwise I’m just a Brainy Smurf at the moment :smiley:

If things proove the problem is software, I’d carefully check kernel versions in the system the discs work, and compare to the systems kernel version in which the discs do not work.

1 Like
12

If I had a system like that - I’d be installing TrueNAS and FreeBSD :heart:

3 Likes
13

Yes , but very short .

Well, I just installed Antix23.1. Its live system can see the new disk, its installer can see the new disk, and when installed and booted it can see the new disk.

I am busy comparing lsmod from Antix with that from MX.
Here are the results
Present in Antix and not in MX

8021q 40960 0
asn1_encoder 16384 1 trusted
dm_crypt 53248 0
encrypted_keys 24576 1 dm_crypt
fb_sys_fops 16384 1 drm_kms_helper
garp 16384 1 8021q
mfd_core 16384 3 intel_pmc_bxt,lpc_ich,amdgpu
mousedev 24576 0
mrp 20480 1 8021q
nls_utf8 16384 1
polyval_clmulni 16384 0
polyval_generic 16384 1 polyval_clmulni
rng_core 20480 2 ath9k,tpm
syscopyarea 16384 1 drm_kms_helper
sysfillrect 16384 1 drm_kms_helper
sysimgblt 16384 1 drm_kms_helper
tee 40960 1 trusted
tpm 81920 1 trusted
trusted 36864 2 encrypted_keys,dm_crypt
xhci_pci_renesas 16384 1 xhci_pci

and
present in MX and not in Antix

ahci 49152 6
amdxcp 12288 1 amdgpu
ansi_cprng 12288 0
async_memcpy 16384 2 raid456,async_raid6_recov
async_pq 16384 2 raid456,async_raid6_recov
async_raid6_recov 20480 1 raid456
async_tx 16384 5 async_pq,async_memcpy,async_xor,raid456,async_raid6_recov
async_xor 16384 3 async_pq,raid456,async_raid6_recov
blake2b_generic 24576 0
bridge 385024 1 br_netfilter
br_netfilter 32768 0
btrfs 2052096 0
cdrom 81920 1 sr_mod
cpufreq_conservative 16384 0
cpufreq_ondemand 20480 12
cpufreq_powersave 16384 0
cpufreq_userspace 16384 0
crc16 12288 2 bluetooth,ext4
crc32c_generic 12288 0
crc64 16384 2 crc64_rocksoft,crc64_rocksoft_generic
crc64_rocksoft 16384 1 t10_pi
crc64_rocksoft_generic 12288 1
crc_t10dif 16384 1 t10_pi
crct10dif_common 12288 3 crct10dif_generic,crc_t10dif,crct10dif_pclmul
crct10dif_generic 12288 0
drbg 49152 1
drm_suballoc_helper 12288 1 amdgpu
ecc 45056 1 ecdh_generic
efivarfs 24576 1
ehci_hcd 110592 1 ehci_pci
ehci_pci 16384 0
evdev 28672 34
fat 102400 1 vfat
fuse 208896 4
hid 176128 2 usbhid,hid_generic
hid_generic 12288 0
ip6t_REJECT 12288 1
ip6t_rt 16384 3
ipt_REJECT 12288 1
jitterentropy_rng 20480 1
libahci 61440 1 ahci
libata 475136 2 libahci,ahci
libcrc32c 12288 5 nf_conntrack,nf_nat,btrfs,nf_tables,raid456
linear 16384 0
loop 36864 0
md_mod 225280 6 raid1,raid10,raid0,linear,raid456,multipath
multipath 16384 0
nf_conntrack 212992 4 xt_conntrack,nf_nat,nf_conntrack_netlink,xt_MASQUERADE
nf_conntrack_netlink 61440 0
nf_defrag_ipv4 12288 1 nf_conntrack
nf_defrag_ipv6 24576 1 nf_conntrack
nf_log_syslog 24576 10
nf_nat 65536 2 nft_chain_nat,xt_MASQUERADE
nfnetlink 20480 4 nft_compat,nf_conntrack_netlink,nf_tables
nf_reject_ipv4 16384 1 ipt_REJECT
nf_reject_ipv6 24576 1 ip6t_REJECT
nf_tables 372736 662 nft_compat,nft_chain_nat,nft_limit
nft_chain_nat 12288 3
nft_compat 20480 121
nft_limit 16384 13
nls_ascii 12288 1
overlay 192512 0
qrtr 57344 4
raid0 24576 0
raid1 57344 0
raid10 77824 0
raid456 200704 0
raid6_pq 122880 4 async_pq,btrfs,raid456,async_raid6_recov
rc_core 73728 2 cec
scsi_common 16384 7 scsi_mod,sd_mod,usb_storage,uas,libata,sg,sr_mod
scsi_mod 327680 6 sd_mod,usb_storage,uas,libata,sg,sr_mod
sd_mod 86016 6
sg 45056 0
sha3_generic 16384 1
sha512_generic 16384 1 sha512_ssse3
sr_mod 28672 0
t10_pi 20480 1 sd_mod
uinput 20480 1
usb_common 20480 3 xhci_hcd,usbcore,ehci_hcd
usbcore 409600 8 xhci_hcd,ehci_pci,usbhid,usb_storage,ehci_hcd,btusb,xhci_pci,uas
vfat 20480 1
vmd 24576 0
watchdog 49152 1 iTCO_wdt
xfrm_algo 16384 1 xfrm_user
xfrm_user 61440 1
xhci_hcd 352256 1 xhci_pci
xor 20480 2 async_xor,btrfs
x_tables 61440 11 xt_conntrack,nft_compat,xt_LOG,xt_tcpudp,xt_addrtype,ip6t_rt,ipt_REJECT,xt_limit,xt_hl,xt_MASQUERADE,ip6t_REJECT
xt_addrtype 12288 6
xt_conntrack 12288 17
xt_hl 12288 22
xt_limit 12288 0
xt_LOG 16384 10
xt_MASQUERADE 16384 1
xt_tcpudp 16384 60

The thing that strikes me is that Antix does not have the ahci module, which is supposed to be essential.
However when I look ad dmesg in Antix , it is using AHCI… the module must have a different name in Antix?

Kernel versions
MX 6.5.0
Antix 6.1.60

1 Like
14

I tried it, connected one disk directly to the SATA cable coming from the PCIE card.
No difference. The BIOS can see the disc, MX can not see it.

I think I am going to have to try a fresh MX install.

4 Likes
15

That’s a valuable information. Deflects even my suspition from the cage :slight_smile:
Would you like to try a “vanilla” Debian live?
Just to see how it behaves…

16

Just noticed this…
So MX has a newer kernel, you may be a victim of a regression in the kernel.

17

Will do.

I just had another idea. it may need an update-initramfs … the disks are setup very early in the boot

1 Like
18

I agree, it would suit that.
I did not do it for that purpose. This is my main work machine. I wanted to try and speed up backups of my workfiles… rsync to a usb drive is slow and a nuisance.
I thought if I had 3 or 4 HDD’s and rotated them using the cage ( it is supposed to hotplug but I have not tried yet), I could make a simple set of rotated backups like we used to do with tapes. And the SATA disks should be faster than usb drives.

I also got one SSD… I will use it as an experimental area… keep my main installs safe

Someday it may be recycled as a NAS.

5 Likes
19

Getting there, downloaded and made the usb drive tonight.

In comparing new Antix install, which works, with existing MX , which does not work
have checked

  • kernel boot parameters
  • modules loaded
  • daemons running
  • udev rules
  • update-initramfs in MX did nothing
    and
    there are no significant differences that I can see.
    so
    it has to be a kernel regression error… MX is 6.5, Antix is 6.1.
    I need to try some older kernels in MX.
3 Likes
20

That would be interesting.