New disk not recognized


Marino

Recommended Posts

Hi,

 

I have a new disk which is noch recognized. On the same cable an other drive (3TB WD Red) works perfektly. Now I have a 12TB HGST 4Kn drive, which is not recognized.

 

Extern in a reader on a Mac the drive works. When I put it in the Unraid-Server it spins up but the server need almost 5 minutes instead of before under one minute to go online, but there is no drive.

 

I've read som about molex and SATA-Adapter which is needed from a few people because it won't spin up, but it is spinning up.

 

It is attatched to a Gigabyte Z87X-UD5H, newest unraid

 

What could this be? I don't see any errors.

 

kind regards

nils

Link to comment
  • Replies 112
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Posted Images

Oh, I totally forgot. I attached the diagostic files.

Archiv.zip

 

The extern Method is an Inatec USB 3.0 to SATA 6.0 DUAL Bay Docking station. The Mac want's to initialize and format. But what format should I use? Unraid should do it itself shouldn't it?

 

I took the WD Red away, because this drive is from another server and it is the easiest way to remove that and use the cables. So I know, that the cables were working before. 

The HGST is running. I can fell the vibrations on it while starting and when it turns. But the head isn't moving.

 

I wanted to test it with molex to SATA adapter, but I haven't got an molex plug on my power supply. Only SATA. But in this case i've read that this could cause problems, when the P3 has voltage, which is not the case when using molex>sata. And this is why the drive will not spin up. But in my case, the drive spins.

 

EDIT:

I just realized, when I connect the 12TB HDD the other unattached drive (320gb) is not recognized either.

I connected the extern Inatec USB to the server with the 12TB in it and it shows under unattached devices, but not green and the format button is grey (picture)

Bildschirmfoto 2018-09-01 um 22.01.12.png

Edited by Marino
Link to comment
5 hours ago, Marino said:

I've read som about molex and SATA-Adapter which is needed from a few people because it won't spin up, but it is spinning up.

 

A recently-manufactured drive may have implemented the 3.3v reset which is new to the SATA standard.  Many older power supplies do not support this.  The drive will spin up (as yours is doing) but is not recognized by the OS because it is in a continual power reset loop.

 

A Molex --> SATA adapter is one way around this as is a USB docking station as both bypass the 3.3v reset function.  One way to test if this is really the issue with this drive is to place a piece of tape (ideally Kapton tape) over pin #3 on the SATA power connector on the drive.  This prevents electrical contact from being made between the SATA power cable and drive on pin 3. 

 

If you don't have Kapton tape, you can use another kind of tape just to test.  Electrical tape is often too thick.  Make sure if you remove the tape after the test that you thoroughly clean off any residue that may remain.

 

This may not be the issue, but, it is easy to test with tape on pin 3 (you can cover pins 1 and 2 as well if you cannot cut a piece small enough for just pin 3.

Link to comment

Simple stuff first. If the disk is spinning you don't need to worry about the +3.3 volt wire. There's a datasheet from HGST. The format button in  Unassigned Devices is only available if you enable destructive mode (Settings -> Unassigned Devices).

 

Are you saying that you are trying to house the new 12 TB disk in an external case, which connects via USB 3.0 and has space for two hard disks? I recommend using a SATA port or eSATA, if you have to, instead. If you really must use USB so that you can, say, detach the disk and use it as an occasional backup, then just buy a case that holds one disk so as to avoid the port multiplier. It isn't being detected properly and there are a lot of resets until it finally gives up. There's no SMART information either.

 

Edited by John_M
Added HGST link
Link to comment

Thank you guys for answering.

 

I was already aware of P3 before ordering, so I ordered the HGST HUH721212ALN600/0F30141. This should be a HGST Ultrastar He12, 12TB, ISE, SATA, 4Kn. The HUH721212ALN600/0F29620 should be the exact same one but in addition to that with P3. I could not even order the second one with P3 here in germany. I compared the numbers. I've got what I ordered. So ideally it has not the P3 feature. 

 

In the datasheedis written, that in case of incompability (PSU with older SATA-Standard), the drive will may not spin up. But oddly enough they don't say that it is definitely not spinning up. So there could be a chance, but with the numbers above the drive should not have this feature.

 

Don't get me wrong John_M. I have 2 more drives on order and need to set them up in the tower. The 2 Bay externally USB 3.0 Station is only for backup reasons and for testing. So I can have a intern HDD without housing (in a little case), do some backup an place it somewhere else. 

 

A few things bothering me:

- Why took the server so long to get online, when the drive is connected

- Even when the P3 is holding up the drive from spinning normal, why is the second unattached device only visible when not using the 12TB drive or when using it on USB?

 

I am testing it right now with preclear. This will take another 6 hours. After that I can test it with the tape, when there is 3.3V and this isolates it, It should start. The only thing is, that I don't have any really thin tape :(

 

Looking for the molex-Adapter of my PSU (Enermax Revolution X't) was unsuccessfull. Maybe there is some after market for that or is it possible to cut the wire only for PIN3 and use some shrinkage tube on it? When the P3 is never in use for SATA this should not be a problem in my opinion. The wire assignment looks the same for molex and SATA, but the SATA in addition has the orange 3.3V wire. Who needs that anyway when not using P3? In my case all wires are black :)

When the drive works, this unraid-server will have 3x12TB and when upgrading I will use 12TB ones. So Molex would only be for adapting to SATA.

This is the SATA-Cable for my Enermax PSU for SATA) it has 5 wires on it instead of 4 (molex). Is the first wire for Pin1,2 and 3 (bridged)?

 

HGST means that SATA REV v.3.2+ or higher is needed. But which part has to be compatible with that? The Datasheet of the Enermax only says "Intel ATX 12V v2.4" and "Intel Hasewell ready with C6/C7", but this is only for CPU. Gigabyte says "SATA 6Gb/s"?!?

 

 

kind regards

nils

 

 

EDIT: I recently asked gigabyte about using 4Kn disks, they should work without problems...

EIDT 2: I forgot to mention. I can read SMART when using it via USB 3.0. Short test was without any problems and 1h of operation time. Data looks good on the first sight.

Edited by Marino
Link to comment

Seems not relate P3 ( new SATA REV ), check your syslog got some odd/bad message

 

Which controller used for that 12TB disk ? Have try onboard / add-on ? I think controller issue rather then 3.3v issue.

 

Sep  1 20:55:28 uni7 kernel: ata5.00: ACPI cmd ef/10:06:00:00:00:00 (SET FEATURES) succeeded
Sep  1 20:55:28 uni7 kernel: ata5.00: ACPI cmd f5/00:00:00:00:00:00 (SECURITY FREEZE LOCK) filtered out
Sep  1 20:55:28 uni7 kernel: ata5.00: ACPI cmd b1/c1:00:00:00:00:00 (DEVICE CONFIGURATION OVERLAY) filtered out

Sep  1 20:55:28 uni7 kernel: ata5.00: ACPI cmd ef/10:06:00:00:00:00 (SET FEATURES) succeeded
Sep  1 20:55:28 uni7 kernel: ata5.00: ACPI cmd f5/00:00:00:00:00:00 (SECURITY FREEZE LOCK) filtered out
Sep  1 20:55:28 uni7 kernel: ata5.00: ACPI cmd b1/c1:00:00:00:00:00 (DEVICE CONFIGURATION OVERLAY) filtered out

Sep  1 20:55:28 uni7 kernel: ata6.00: ACPI cmd ef/10:06:00:00:00:00 (SET FEATURES) succeeded
Sep  1 20:55:28 uni7 kernel: ata6.00: ACPI cmd f5/00:00:00:00:00:00 (SECURITY FREEZE LOCK) filtered out
Sep  1 20:55:28 uni7 kernel: ata6.00: ACPI cmd b1/c1:00:00:00:00:00 (DEVICE CONFIGURATION OVERLAY) filtered out

 

Sep  1 20:55:28 uni7 kernel: ata13: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata9: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata11: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata10: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata8: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata14: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata7: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata12: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata13: COMRESET failed (errno=-16)
Sep  1 20:55:28 uni7 kernel: ata9: COMRESET failed (errno=-16)
Sep  1 20:55:28 uni7 kernel: ata10: COMRESET failed (errno=-16)
Sep  1 20:55:28 uni7 kernel: ata11: COMRESET failed (errno=-16)
Sep  1 20:55:28 uni7 kernel: ata8: COMRESET failed (errno=-16)
Sep  1 20:55:28 uni7 kernel: ata14: COMRESET failed (errno=-16)
Sep  1 20:55:28 uni7 kernel: ata12: COMRESET failed (errno=-16)
Sep  1 20:55:28 uni7 kernel: ata7: COMRESET failed (errno=-16)
Sep  1 20:55:28 uni7 kernel: ata13: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata12: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata11: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata8: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata10: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata9: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata14: link is slow to respond, please be patient (ready=0)
Sep  1 20:55:28 uni7 kernel: ata7: link is slow to respond, please be patient (ready=0)

Edited by Benson
Link to comment

Hi Benson,

 

I've got the Gigabyte Z87X-UD5H without any additional Controller.

So I have 6 SATA-Ports with Chipset (Intel Z87 Express Chipset) and 4 SATA Ports with Marvell 88SE9230). 

 

The 12TB disk is the second on the Marvell-Controller. The first one is an unattached disk and this is the disk which disappears when trying to connect the 12TB on the second port.

Unfortunately this controller seems to have problems with virtualization. But because this is my old PC and it has an K-CPU (Intel 4770K) and the K-CPU's don't support virtualization. This is grey in BIOS. This is why i couldn't use my graphic card for VM.

 

The Intel-Controller is fully used with parity, 4 disks and cache. The parity should be replaced by this one. So the best way to test would be to disconnect the cache in order to have the array untouched.

 

But even when the Marvell-Controller could cause problems, why is the unattached device on it working for months? This equals not the error pattern which is described here.?!? 

 

Link to comment

Assuming that the controller has some problems then I would lose 4 Ports to use. 

 

Could I take the cache off and place the 12TB here, then make a parity swap and replace it with another disk (parity) on port 1? I could now (when working) use port 6 and after parity swap I would connect it to port 1. Is this problematically or is it possible to stop the array and assign a working disk to another port?

Link to comment

Looks like all the Marvell controller ports stopped working with that disk connected, Marvell controllers are not recommended for unRAID but if it was working OK with just the other disk you could try swapping one of the smaller disks connected to the Intel controller there and connect the new disk on the Intel port.

Link to comment

So swapping a disk is okay? Just swap to another port and reassign it in unraid?

 

Sooner or later there will be only 12TB disks, So for 7+ disks I'll need another controller, if thats the problem.

 

EDIT:

The disk which is working on that is an 250GB 2,5" WD Blue. Maybe that does have not the same effect als really large ones?!? 

Edited by Marino
Link to comment
2 hours ago, Marino said:

The first one is an unattached disk and this is the disk which disappears when trying to connect the 12TB on the second port.

 

18 hours ago, Marino said:

EDIT:

I just realized, when I connect the 12TB HDD the other unattached drive (320gb) is not recognized either.

I connected the extern Inatec USB to the server with the 12TB in it and it shows under unattached devices, but not green and the format button is grey (picture)

 

1 hour ago, Marino said:

I just saw that my Cache Disk (SSD on Intel Port 6) isn't recognized anymore ether.

 

As 12TB drive no problem when connect thr USB bridge, so I will assume disk haven't issue. And your system was all normal before adding this disk.

 

When connect the 12TB disk, many unexpected come out. What I think is some AHCI with 4Kn / large drive compatibility problem. It may not HW related. Whatever Intel or Marvell controller, those are under AHCI. Anyway you need test with different change.

 

Pls stop array auto start and check "log" & "boot time" for every change.

 

1. If 12TB disk disconnect, does system resume normal and no those error in log ?

2. If 12TB disk connect ( Intel or Marvell ), does same error could reproduce.

3. If above all yes, would you try change the BIOS AHCI to IDE and check any different ?

 

Edited by Benson
Link to comment

I missunderstood preclear, when It shows almost 100%. This may take a while, because it is now in Step 2/5 (zeroing). What I'll test next is when preclear is finished to connect it instead of the cache on the Intel. And I should take a look into the BIOS, because of virtualisazion and marvell. Maybe it is a controller-problem with marvell, because 4Kn should run (as the gigabyte service said). 

 

When nothing helps, I disconnect P3 in any way. But I will keep you up to date, even when it takes a "few" hours to finish preclearing.

Edited by Marino
Link to comment

Mhhh. PreClear is almost ready, but now it is stucked. "unRAID's signature on the MBR is valid" and in preview ""Step 5 of 5 - Post-Read in progress ...".

 

The Elapsed time is "frozen" since about 9 hours and the disk is getting cooler. From 43 to 44°C while writing zeros to 38°C.

 

How can I see, if preclear is doing something or not?

Bildschirmfoto 2018-09-03 um 11.24.13.png

Bildschirmfoto 2018-09-03 um 11.24.04.png

Edited by Marino
Link to comment

The log looked like I never startet so I decided to end this thing an start it from scratch. I was lucky. PreClear could resume the clearing and is now at 0% of Step 5. I wish I'd known that before than Preclear would be done in a few hours. So it'll take another 13 to 15 hours and I could have saved some time for researching this problem :(

Edited by Marino
Link to comment

PreClear is Done. I've connect the drive to the Intel-Controller and it works immediately. So I can blame the Marvell-Controller.

 

I just wantet to make a parity swap with copying the parity disk like the parity swap procedere in the wiki describes. But sadly there is no copy-button so my only chance is to compute the parity again. Because of 12TB, it'll take 35 hours. Copying 4TB would have been much faster. :(

 

EDIT:

I wonder if it is possible to upgrade the marvell-firmware on the Gigabyte-Mainboard.

 

Edited by Marino
Link to comment
18 minutes ago, Marino said:

I just wantet to make a parity swap with copying the parity disk like the parity swap procedere in the wiki describes. But sadly there is no copy-button so my only chance is to compute the parity again. Because of 12TB, it'll take 35 hours. Copying 4TB would have been much faster. :(

If there's no copy button you're doing something wrong.

 

18 minutes ago, Marino said:

I wonder if it is possible to upgrade the marvell-firmware on the Gigabyte-Mainboard.

Much better to not use it anymore, lots of problems with Marvell controllers and unRAID, get an LSI instead.

Link to comment

You're welcome. I have to thank you helping me, solving this problem. So: "Thank you guys for your awesome help!". We can blame marvell for this. After parity computing, adding two more 12TB disks and copying from the old disks, I'll only have 2 data and 1 parity drives and a cache. So, next week as it seems. Which gives me 2 more ports on Intel. Saying this, I have to fill about 15TB of Data before I've to use this. 

 

When I need more ports, I'd go with an extra controller instead of using marvell.

 

@johnnie.black

This is postet every time: https://wiki.unraid.net/The_parity_swap_procedure

Maybe I missunderstood, but in #13 is written that there should be a copy-button to copy the parity data to the new drive. Am i wrong?

 

I am doing this upgrade to get 3 4TB drives out of this server. I had problems with my old unraid-server, which has one parity and 8 3TB drives in total. I have to reconstruct one drive, because of cable issues. 2 of the 4TB drives will build a dual parity and the third one is for rebuilding and not touching the drive which was causing the problem.

 

Therefore I have to swap one parity drive and add one more to it (at the end). When the swap procedure is wrong, what is the right way to do it without computing parity at least for one drive. 

 

We can discuss this in my other thread if thats the better place for that. 

 

While I tried to make the swap (wiki) I managed to make the parity invalid. In no case this should happen on the other server. Because I need for reconstruction a 4TB drive in a 3TB-enviroment. Before I can do this, an upgrade of parity (3>4TB) is needed. When the 3TB parity gets invalid in either way, I've got some problems reconstructing the drive which is offline!

 

 

This is my other thread, where I have to make a parity-swap when this server is ready, without loosing parity, because I need it to reconstruct a disk.

 

Link to comment

I uploaded this in the other thread before. 

 

This is my other server. When I am done here, I'll get 3 4TB drives. 

 

Disk7 has to be rebuild. 

2 of the 4TB drives should be a dual parity and one is for replacing the Disk7. 

 

Because they're 4TB I have to swap one parity before rebuilding on a 4TB drive.

350155589_Bildschirmfoto2018-08-20um19_18_20.png.8d845268b6e4b7529f04f002a3d36c5e-2.png

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.