Disk wrong : device is disabled

[Darkel]

Hello,

 

My Array can't start this day. There is an error on a disk, message overlay is : device is disabled : contents emulated

I use this disk as an entire disk "passthrough" in a VM using option /dev/disk/by-id/ata-Samsung_SSD_860_QVO_1TB_S4CZNF0N613815T so the disk cannot be emulated in the array.

 

I change SATA cable, used another SATA port but error remains. I did a smart short test and a smart long test but no errors.

 

NB: I use an entire disk to my VM because I had bad performance when using my SSD from Array. My SSD can perform 500read/ 500write mb/s when used outside Unraid but if I use it in a Array I had random performance, it can be up to 500writes mb/s (write) but in several times it cannot exceed 80mb/s (write). I used crystal hdd benchmark.

 

Feel free to help this

Thank you.

tower-syslog-20220310-1607.zip tower-smart-20220310-1706.zip

Edited March 11, 2022 by Darkel

[trurl]

Instead of syslog and smart we always prefer diagnostics. Diagnostics includes syslog, SMART for all attached disks, and lots of other information about your hardware and configuration. I seldom even look at syslog without diagnostics, they help me understand what needs to be looked for in syslog among other things.

 

Please attach diagnostics to your NEXT post in this thread, and clarify what you have done.

 

This disk was originally in the array? Did you New Config and rebuild parity without it? If not, then it is still in the array as far as Unraid is concerned.

 

[Darkel]

Hello,

Thanks for this quick reply.

 

I dont have, I don't use parity disk or cache disk.

Disk was originally in the array, yes. I used it as part of the array for a while as an xfs disk and I created and mounted a vdisk to my VM. When I started to explore my disk performance I delete vdisk and mounted the entire disk (using /dev/disk/by-id/) to my windows VM and formated it using ntfs. I think that the issue first occured when I reboot, I think it's the point. I had folders in /mnt/user/domains, iso, etc. Okay I understand now, Unraid cannot find back all its "SHARE" folders. 

 

I think it must be impossible to recover any thing on the array. I had another vdisk on /mnt/user/domain/disk2 I believed that if I put something on /mnt/user/domain/disk# it should be physically on the #disk but if I think in the array logic I think it should be lost, aren't you?

 

I won't cry a lot about situation because there is nothing important, it's just about time to recreate all my vm back. (I use Nextcloud on another server to backup all my important files). But how can I recreate a new array? And is there something to avoid if I want to passthrough an entire disk?

 

Thank you.

tower-diagnostics-20220310-2039.zip

Edited March 10, 2022 by Darkel

[trurl]

3 hours ago, Darkel said:

/mnt/user/domain/disk#

This is a path in the user share named "domain", could be on any disk depending on settings for user share domain.

 

If you want to specify a disk path, that would be /mnt/disk#

3 hours ago, Darkel said:

how can I recreate a new array

New Config will let you assign disks however you want and when you start the array those assignments will be recorded. Parity will be built on any disk assigned to any parity slot.

 

3 hours ago, Darkel said:

something to avoid if I want to passthrough an entire disk?

Don't assign it to the array.

[trurl]

Also, SSDs assigned to the array can't be trimmed. Better if SSDs are assigned to pools, but you must have at least one array disk even if you don't use it. Some people use another USB flash drive for this purpose. Do you intend to have any NAS functionality at all?

[Darkel]

Hello,

 

I performed a new config and then I only included a USB disk (128gb), the array will be used only for internal unraid use, like mounting operating system for a new vm, mounting driver iso-file to a vm and hosting rom file for gpu. I don't use NAS functionality, I use Unraid as gaming usage, I split my CPU core into 2 VM, split my 16GB RAM (would like to go to 32GB), dedicate 1 SSD per VM and I have 2 GPU, one for each VM,. I love this way

 

Thank to you I understand all the story now. Thank you again, my brain feels better now.

Edited March 11, 2022 by Darkel

[Darkel]

I missed it,

 

Do you know the reason of these "parts" ?

 

root@Tower:/dev/disk/by-id# ls -al
total 0
drwxr-xr-x 2 root root 680 Mar 11 12:31 ./
drwxr-xr-x 8 root root 160 Mar 11 11:30 ../
lrwxrwxrwx 1 root root   9 Mar 11 12:31 ata-Samsung_SSD_860_QVO_1TB_S4CZNF0N613815T -> ../../sdb
lrwxrwxrwx 1 root root  10 Mar 11 12:31 ata-Samsung_SSD_860_QVO_1TB_S4CZNF0N613815T-part1 -> ../../sdb1
lrwxrwxrwx 1 root root  10 Mar 11 12:31 ata-Samsung_SSD_860_QVO_1TB_S4CZNF0N613815T-part2 -> ../../sdb2
lrwxrwxrwx 1 root root  10 Mar 11 12:31 ata-Samsung_SSD_860_QVO_1TB_S4CZNF0N613815T-part3 -> ../../sdb3
lrwxrwxrwx 1 root root  10 Mar 11 12:31 ata-Samsung_SSD_860_QVO_1TB_S4CZNF0N613815T-part4 -> ../../sdb4
lrwxrwxrwx 1 root root  10 Mar 11 12:31 ata-Samsung_SSD_860_QVO_1TB_S4CZNF0N613815T-part5 -> ../../sdb5
lrwxrwxrwx 1 root root   9 Mar 11 11:30 ata-SanDisk_SDSSDHII480G_154076400609 -> ../../sdd
lrwxrwxrwx 1 root root  10 Mar 11 11:30 ata-SanDisk_SDSSDHII480G_154076400609-part1 -> ../../sdd1
lrwxrwxrwx 1 root root   9 Mar 11 11:30 ata-SanDisk_SDSSDHP128G_134362401271 -> ../../sdc
lrwxrwxrwx 1 root root  10 Mar 11 11:30 ata-SanDisk_SDSSDHP128G_134362401271-part1 -> ../../sdc1
lrwxrwxrwx 1 root root  10 Mar 11 11:30 ata-SanDisk_SDSSDHP128G_134362401271-part2 -> ../../sdc2
lrwxrwxrwx 1 root root  10 Mar 11 11:30 ata-SanDisk_SDSSDHP128G_134362401271-part3 -> ../../sdc3
lrwxrwxrwx 1 root root  10 Mar 11 11:30 ata-SanDisk_SDSSDHP128G_134362401271-part4 -> ../../sdc4
lrwxrwxrwx 1 root root  10 Mar 11 11:30 ata-SanDisk_SDSSDHP128G_134362401271-part5 -> ../../sdc5
lrwxrwxrwx 1 root root   9 Mar 11 11:30 usb-SanDisk_Ultra_4C530000160311204483-0:0 -> ../../sda
lrwxrwxrwx 1 root root  10 Mar 11 11:30 usb-SanDisk_Ultra_4C530000160311204483-0:0-part1 -> ../../sda1
lrwxrwxrwx 1 root root   9 Mar 11 12:22 usb-SanDisk_Ultra_4C530001090805121345-0:0 -> ../../sde
lrwxrwxrwx 1 root root  10 Mar 11 12:22 usb-SanDisk_Ultra_4C530001090805121345-0:0-part1 -> ../../sde1
lrwxrwxrwx 1 root root   9 Mar 11 11:30 wwn-0x5001b44a543485f7 -> ../../sdc
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44a543485f7-part1 -> ../../sdc1
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44a543485f7-part2 -> ../../sdc2
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44a543485f7-part3 -> ../../sdc3
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44a543485f7-part4 -> ../../sdc4
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44a543485f7-part5 -> ../../sdc5
lrwxrwxrwx 1 root root   9 Mar 11 11:30 wwn-0x5001b44eeb4047e1 -> ../../sdd
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44eeb4047e1-part1 -> ../../sdd1
lrwxrwxrwx 1 root root   9 Mar 11 12:31 wwn-0x5002538e90618ff6 -> ../../sdb
lrwxrwxrwx 1 root root  10 Mar 11 12:31 wwn-0x5002538e90618ff6-part1 -> ../../sdb1
lrwxrwxrwx 1 root root  10 Mar 11 12:31 wwn-0x5002538e90618ff6-part2 -> ../../sdb2
lrwxrwxrwx 1 root root  10 Mar 11 12:31 wwn-0x5002538e90618ff6-part3 -> ../../sdb3
lrwxrwxrwx 1 root root  10 Mar 11 12:31 wwn-0x5002538e90618ff6-part4 -> ../../sdb4
lrwxrwxrwx 1 root root  10 Mar 11 12:31 wwn-0x5002538e90618ff6-part5 -> ../../sdb5

 

[trurl]

56 minutes ago, Darkel said:

the reason of these "parts"

Unraid will not have created those. Are these not new disks?

[Darkel]

ata-Samsung_SSD_860_QVO_1TB_S4CZNF0N613815T

ata-SanDisk_SDSSDHII480G_154076400609

These two were originaly in the first array I created.

For now ata-Samsung_SSD_860_QVO_1TB_S4CZNF0N613815T is used as entire disk and the second one is not used and is not in the array.

 

ata-SanDisk_SDSSDHP128G_134362401271

was added to the first array but used as "entire disk" in a VM (using path : /dev/disk/by-id/ata-SanDisk_SDSSDHP128G_134362401271)

 

usb-SanDisk_Ultra_4C530000160311204483-0:0 must be the disk where unraid is installed

usb-SanDisk_Ultra_4C530001090805121345-0:0 and this one used for the new and current array

 

And these one, I don't know what are they:

lrwxrwxrwx 1 root root   9 Mar 11 11:30 wwn-0x5001b44a543485f7 -> ../../sdc
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44a543485f7-part1 -> ../../sdc1
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44a543485f7-part2 -> ../../sdc2
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44a543485f7-part3 -> ../../sdc3
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44a543485f7-part4 -> ../../sdc4
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44a543485f7-part5 -> ../../sdc5
lrwxrwxrwx 1 root root   9 Mar 11 11:30 wwn-0x5001b44eeb4047e1 -> ../../sdd
lrwxrwxrwx 1 root root  10 Mar 11 11:30 wwn-0x5001b44eeb4047e1-part1 -> ../../sdd1
lrwxrwxrwx 1 root root   9 Mar 11 12:31 wwn-0x5002538e90618ff6 -> ../../sdb
lrwxrwxrwx 1 root root  10 Mar 11 12:31 wwn-0x5002538e90618ff6-part1 -> ../../sdb1
lrwxrwxrwx 1 root root  10 Mar 11 12:31 wwn-0x5002538e90618ff6-part2 -> ../../sdb2
lrwxrwxrwx 1 root root  10 Mar 11 12:31 wwn-0x5002538e90618ff6-part3 -> ../../sdb3
lrwxrwxrwx 1 root root  10 Mar 11 12:31 wwn-0x5002538e90618ff6-part4 -> ../../sdb4
lrwxrwxrwx 1 root root  10 Mar 11 12:31 wwn-0x5002538e90618ff6-part5 -> ../../sdb5

 

Edited March 11, 2022 by Darkel

[trurl]

3 hours ago, Darkel said:

must be the disk where unraid is installed

Unraid is not installed on a disk. The archives of the OS are unpacked fresh from the flash drive into RAM at boot, and it runs completely in RAM. Think of it as firmware.

 

4 hours ago, trurl said:

Are these not new disks?

Doesn't seem like you answered this, or at least I didn't understand your answer. The point of my question was that those partitions must have been created when these disks were previously used in some other OS.

[trurl]

21 hours ago, Darkel said:

something to avoid if I want to passthrough an entire disk?

Also don't assign to a pool. Unraid must not even try to work with the disk, even as an Unassigned Device, if you are going to pass it to a VM.

[Darkel]

On 3/11/2022 at 6:00 PM, trurl said:

The point of my question was that those partitions must have been created when these disks were previously used in some other OS.

Oh okay, I understand and this is my answer : these "parts" could be in fact the partitions that has been created by windows 10, because windows create a recovery partition, (1) a system partition (2), a data partition (3), another recovery partition (4) and I have an unallocated partition (5), so it must be the answer! Thank to you

 

So I understand that Unraid automatically make a mounting point of each partition. But when a disk is used in the array, there is no part because, the entire disk is used inside the array.

 

I used this disk as part of the array. But I also used it as entire disk for a VM:

lrwxrwxrwx 1 root root   9 Mar 11 11:30 ata-SanDisk_SDSSDHII480G_154076400609 -> ../../sdd
lrwxrwxrwx 1 root root  10 Mar 11 11:30 ata-SanDisk_SDSSDHII480G_154076400609-part1 -> ../../sdd1

That's why there is -part1 disk in addition of the first line.

 

Is it correct? 😄

[trurl]

4 hours ago, Darkel said:

lrwxrwxrwx 1 root root   9 Mar 11 11:30 ata-SanDisk_SDSSDHII480G_154076400609 -> ../../sdd
lrwxrwxrwx 1 root root  10 Mar 11 11:30 ata-SanDisk_SDSSDHII480G_154076400609-part1 -> ../../sdd1

sdd is the device

sdd1 is the first (and only) paritition on the device

 

4 hours ago, Darkel said:

I used this disk as part of the array. But I also used it as entire disk for a VM

...

Is it correct?

 

If you want to pass the disk to a VM, don't let Unraid use it at all, not as any array disk, not as a pool disk, and not as an Unassigned Device.