Disk Spin up

i'm thinking we could work on this again when rc3 is released.

On 1/21/2022 at 12:11 PM, SimonF said:

No worries I can re test if required.

Have you updated to RC3?  If not then please do that and let's work on this again.

1 hour ago, dlandon said:

Have you updated to RC3?  If not then please do that and let's work on this again.

Yes have both my test systems on rc3 now. I will try test tomorrow.

Let me know how it goes and we'll pick back up if you still have the issues.

Still is producing the same issue on RC3. SDD is the parity drive. UD is 2022.03.11

 

JFYI Disks are SAS.

 

Mar 13 07:56:51 Tower emhttpd: spinning down /dev/sdb
Mar 13 07:56:52 Tower emhttpd: read SMART /dev/sdb
Mar 13 07:58:30 Tower kernel: usb 3-6: new high-speed USB device number 6 using xhci_hcd
Mar 13 07:58:30 Tower kernel: usb-storage 3-6:1.0: USB Mass Storage device detected
Mar 13 07:58:30 Tower kernel: scsi host10: usb-storage 3-6:1.0
Mar 13 07:58:31 Tower usb_manager: Info: rc.usb_manager usb_add USB_2.0_Flash_Disk_2006122205051671 /dev/bus/usb/003/006 003 006
Mar 13 07:58:31 Tower usb_manager: Info: rc.usb_manager Autoconnect No Mapping found USB_2.0_Flash_Disk_2006122205051671 /dev/bus/usb/003/006 003 006 port 3-6
Mar 13 07:58:31 Tower kernel: scsi 10:0:0:0: Direct-Access     USB 2.0  Flash Disk       2.00 PQ: 0 ANSI: 2
Mar 13 07:58:31 Tower kernel: sd 10:0:0:0: Attached scsi generic sg12 type 0
Mar 13 07:58:31 Tower kernel: sd 10:0:0:0: [sdm] 3987456 512-byte logical blocks: (2.04 GB/1.90 GiB)
Mar 13 07:58:31 Tower kernel: sd 10:0:0:0: [sdm] Write Protect is off
Mar 13 07:58:31 Tower kernel: sd 10:0:0:0: [sdm] Mode Sense: 0b 00 00 08
Mar 13 07:58:31 Tower kernel: sd 10:0:0:0: [sdm] No Caching mode page found
Mar 13 07:58:31 Tower kernel: sd 10:0:0:0: [sdm] Assuming drive cache: write through
Mar 13 07:58:31 Tower kernel: sdm: sdm1
Mar 13 07:58:31 Tower kernel: sd 10:0:0:0: [sdm] Attached SCSI removable disk
Mar 13 07:58:33 Tower unassigned.devices: Disk with ID 'Flash_Disk_2006122205051671 (sdm)' is not set to auto mount.
Mar 13 07:58:34 Tower emhttpd: USB_2.0_Flash_Disk_2006122205051671-0:0 (sdm) 512 3987456
Mar 13 07:58:41 Tower kernel: sd 2:0:1:0: attempting task abort!scmd(0x000000000a171de2), outstanding for 7038 ms & timeout 7000 ms
Mar 13 07:58:41 Tower kernel: sd 2:0:1:0: [sdd] tag#2559 CDB: opcode=0x12 12 01 00 00 fe 00
Mar 13 07:58:41 Tower kernel: scsi target2:0:1: handle(0x000e), sas_address(0x5000c500626c01e9), phy(2)
Mar 13 07:58:41 Tower kernel: scsi target2:0:1: enclosure logical id(0x5003048011c7eb00), slot(1) 
Mar 13 07:58:41 Tower kernel: sd 2:0:1:0: task abort: SUCCESS scmd(0x000000000a171de2)
Mar 13 07:58:41 Tower emhttpd: read SMART /dev/sdd
Mar 13 07:58:41 Tower emhttpd: read SMART /dev/sdm

 

 

7 hours ago, SimonF said:

Still is producing the same issue on RC3. SDD is the parity drive. UD is 2022.03.11

Can you post your diagnostics?  We've added some additional information to the diagnostics that might help me track this down.

Diags Attached.

tower-diagnostics-20220313-1613.zip

I see two unassigned disks.  sdb and sdi.  I know UD has them right because I can see UD trying to auto mount both of them, but they are not set for auto mount.

 

What does the dashboard show?  What does UD show on the UD page?

6 minutes ago, dlandon said:

I see two unassigned disks.  sdb and sdi.  I know UD has them right because I can see UD trying to auto mount both of them, but they are not set for auto mount.

 

What does the dashboard show?  What does UD show on the UD page?

Before pluging in Device. Parity is spun down.

 

 

Adding Device

 

 

Diags taken at 2245.

tower-diagnostics-20220313-2245.zip

So the unassigned devices are staying straightened out now.  The issue is that the parity disk spins up when an unassigned disk is installed?

3 minutes ago, dlandon said:

So the unassigned devices are staying straightened out now.  The issue is that the parity disk spins up when an unassigned disk is installed?

Yes that is correct.

So we have made some progress.  Now do this:

• Unplug the Dev 4 device.
• Go the the UD Settings and set the "Debug Log Level" to "Udev and Hotplug".
• Plug the Dev 4 device back in.
• Post diagnostics again.

 

This will log the hot plug events UD tracks.  I'm looking to see if a hot plug event is triggering a spin up of the parity disk by mistake.

16 minutes ago, dlandon said:

This will log the hot plug events UD tracks.  I'm looking to see if a hot plug event is triggering a spin up of the parity disk by mistake.

Let me know if you find anything and I will take a look in the morning.

tower-diagnostics-20220313-2317.zip

Eveything looks good.  Next thing to try.  Remove the following plugins:

• DevPack.plg.
• sas-spindown.plg.
• snapshots.plg.
• usb_manager.plg.
• usb_manager_usbip_addon.plg.
• unassigned.devices.plg.

Then run your test.  Let's see if one of the plugins is causing the issue.  This will let us know if it's a plugin or Unraid.

 

With UD uninstalled you won't see the unassigned disk you install in the Unassigned Devices page or the Dashboard.  UD is catching the udev events and triggering Unraid to re-arrange unassigned devices.  Unraid won't do it without UD.

I was going to ask why is Dev3 missing

 

I have removed all of the plugins except SAS spin down other wises all devices will spinup.

 

but found sdf is the parity.

 

 

Rebooted and then added USB flash. All disks where spinning.

 

 

tower-diagnostics-20220314-0734.zip

Edited March 14, 2022 by SimonF

With all disks spundown removed and added flash

 

Mar 14 07:36:22 Tower kernel: usb 3-6: USB disconnect, device number 6
Mar 14 07:36:33 Tower kernel: usb 3-6: new high-speed USB device number 7 using xhci_hcd
Mar 14 07:36:33 Tower kernel: usb-storage 3-6:1.0: USB Mass Storage device detected
Mar 14 07:36:33 Tower kernel: scsi host10: usb-storage 3-6:1.0
Mar 14 07:36:34 Tower kernel: scsi 10:0:0:0: Direct-Access     USB 2.0  Flash Disk       2.00 PQ: 0 ANSI: 2
Mar 14 07:36:34 Tower kernel: sd 10:0:0:0: Attached scsi generic sg12 type 0
Mar 14 07:36:34 Tower kernel: sd 10:0:0:0: [sdm] 3987456 512-byte logical blocks: (2.04 GB/1.90 GiB)
Mar 14 07:36:34 Tower kernel: sd 10:0:0:0: [sdm] Write Protect is off
Mar 14 07:36:34 Tower kernel: sd 10:0:0:0: [sdm] Mode Sense: 0b 00 00 08
Mar 14 07:36:34 Tower kernel: sd 10:0:0:0: [sdm] No Caching mode page found
Mar 14 07:36:34 Tower kernel: sd 10:0:0:0: [sdm] Assuming drive cache: write through
Mar 14 07:36:34 Tower kernel: sdm: sdm1
Mar 14 07:36:34 Tower kernel: sd 10:0:0:0: [sdm] Attached SCSI removable disk

 

Re-installed UD

 

 

Unplugged Flash

 

Mar 14 07:42:41 Tower unassigned.devices: Udev: Refresh udev info for /dev/disk/by-id/usb-SanDisk_Cruzer_Blade_4C530000190606214410-0:0.
Mar 14 07:42:41 Tower unassigned.devices: Udev: Refresh udev info for /dev/disk/by-id/usb-SanDisk_Cruzer_Blade_4C530000190606214410-0:0-part1.
Mar 14 07:42:41 Tower unassigned.devices: Udev: Refresh udev info for /dev/disk/by-id/usb-SanDisk_Cruzer_Blade_4C530000190606214410-0:0-part1.
Mar 14 07:42:41 Tower unassigned.devices: Udev: Refresh udev info for /dev/disk/by-id/scsi-35000cca027baa9a8.
Mar 14 07:42:41 Tower unassigned.devices: Udev: Refresh udev info for /dev/disk/by-id/scsi-35000cca027baa9a8-part1.
Mar 14 07:42:41 Tower unassigned.devices: Udev: Refresh udev info for /dev/disk/by-id/usb-USB_2.0_Flash_Disk_2006122205051671-0:0.
Mar 14 07:42:41 Tower unassigned.devices: Udev: Refresh udev info for /dev/disk/by-id/usb-USB_2.0_Flash_Disk_2006122205051671-0:0-part1.
Mar 14 07:43:06 Tower kernel: usb 3-6: USB disconnect, device number 7
Mar 14 07:43:06 Tower unassigned.devices: Reload: A udev 'remove disk' initiated a reload of udev info.
Mar 14 07:43:06 Tower unassigned.devices: Updating udev information...
Mar 14 07:43:06 Tower unassigned.devices: Udev: Update udev info for /dev/disk/by-id/usb-USB_2.0_Flash_Disk_2006122205051671-0:0.
Mar 14 07:43:06 Tower unassigned.devices: Udev: Update udev info for /dev/disk/by-path/pci-0000:00:14.0-usb-0:6:1.0-scsi-0:0:0:0.
Mar 14 07:43:08 Tower unassigned.devices: Processing Hotplug event...

 

This was devs.ini before and after unplug. Parity drive added to devs.ini

 

root@Tower:/usr/local/emhttp/state# cat devs.ini
["dev1"]
name="dev1"
id="SanDisk_Cruzer_Blade_4C530000190606214410-0:0"
device="sdb"
transport="usb"
sectors="30031872"
sector_size="512"
rotational="1"
spundown="0"
temp="*"
numReads="148"
numWrites="0"
["dev2"]
name="dev2"
id="HUS724030ALS640_P8K9ND8W_35000cca027baa9a8"
device="sdi"
transport="scsi"
sectors="5860533168"
sector_size="512"
rotational="1"
spundown="1"
temp="*"
numReads="108"
numWrites="0"
root@Tower:/usr/local/emhttp/state# cat devs.ini
["dev1"]
name="dev1"
id="SanDisk_Cruzer_Blade_4C530000190606214410-0:0"
device="sdb"
transport="usb"
sectors="30031872"
sector_size="512"
rotational="1"
spundown="0"
temp="*"
numReads="148"
numWrites="0"
["dev2"]
name="dev2"
id="HUS724030ALS640_P8K9ND8W_35000cca027baa9a8"
device="sdi"
transport="scsi"
sectors="5860533168"
sector_size="512"
rotational="1"
spundown="1"
temp="*"
numReads="108"
numWrites="0"
["dev3"]
name="dev3"
id="ST4000NM0023_35000c500626c01eb"
device="sdf"
transport="scsi"
sectors="7814037168"
sector_size="512"
rotational="1"
spundown="0"
temp="*"
numReads="36"
numWrites="0"
root@Tower:/usr/local/emhttp/state# 

 

4 hours ago, SimonF said:

I was going to ask why is Dev3 missing

Missing from where?

 

4 hours ago, SimonF said:

I have removed all of the plugins except SAS spin down other wises all devices will spinup.

Unless we get to a bare metal Unraid we can't determine where the issue is.

 

4 hours ago, SimonF said:

but found sdf is the parity.

I don't see that in the diagnostics.  Only dev1 and dev2 are unassigned.  I find it hard to believe that dynamix is creating unassigned disks out of the clear blue.  It uses the vars dev array to display the unassigned disks.

 

3 hours ago, SimonF said:

This was devs.ini before and after unplug. Parity drive added to devs.ini

I need the diagnostics so I can see the udev and hotplug activity.

 

You need to slow down and do one thing at a time.  It's hard for me to follow when you do things and just give me snippets of logs and files.  I can't follow and there are some things I want to try at each step.

 

So let's get back to the issue with the parity disk showing as unassigned.  How did you create that issue, and is it repeatable?  Create it again if you can, then give me diagnostics and pause there so I can have you do some things.

Here you go

 

Dev3 was missing, Always was dev 1, 2 and 4 never 3.

tower-diagnostics-20220314-1226.zip

2 minutes ago, SimonF said:

Here you go

 

Dev3 was missing, Always was dev 1, 2 and 4 never 3.

tower-diagnostics-20220314-1226.zip

There's an issue with the assignment of the devX designations.  They aren't consistent and can change.  This is something I want to solve, but we haven't found a reasonable solution.  It's impractical for Unraid to track all of the devices ever installed in a server by serial number and devX.

 

Let's move to the parity being shown as unassigned.  Make that happen and give me diagnostics.

Install sas plugin, sleep disks and plugin flash

tower-diagnostics-20220314-1301.zip

UD does not see sdf (parity disk) as a hot plugged device.  I think we might have an Unraid issue.  Can you boot in safe mode, create the issue and then post diagnostics.  I can then present this as an Unraid issue with all plugins out of the way to be sure there is not a plugin problem.

I cannot test in Safe mode as system will not handle SAS drives correctly without the plugin.

 

I have removed UD  amd Issue only occurs if drives are spundown.

 

if I run /usr/local/sbin/emcmd 'cmdHotplug=apply' the following appears after connecting flash drive. Do you know process cmdHotplug is using?

 

So the only plugin installed is the sas drive plugin?

 

3 minutes ago, SimonF said:

Do you know process cmdHotplug is using

No.

 

I will need you to do this with the minimum of plugins and a diagnostic.

Sorry thought I had posted this.

 

Used Smartctl -s standby,now to spin down disks.

 

and ran emcmd ....

 

 

tower-diagnostics-20220314-1839.zip

Edited March 14, 2022 by SimonF

Thank you.  Unraid is in fact adding the parity disk to the unassigned disks,  I'll report this and see if we can get a resolution.