emhttpd: problem getting id and BTRFS error

[Markland]

I have a problem with my unraid set up. It has to deal with the ssds in my system that I am using for my cache drive. The error I believe started primarily when I started using auto backup but I believe the errors were happening all along since I upgraded my unraid set up to include a Netapp DS2246 for my hhd and ssd. Its all being controlled by a LSI 9200-16e.

I can reliable generate the issue by causing the auto back up tool to auto back up and bring the cache drive down. Once the back up is done I get a constant disk write error and all the apps trying to use my cache disk which is the disk in question cant access it but its still actively mounted to the array. 

 

The errors when the array is not mounted yet:

Nov 2 19:54:46 GLaDOS emhttpd: device /dev/sdq problem getting id
Nov 2 19:54:46 GLaDOS emhttpd: device /dev/sdo problem getting id
This is said constantly.

After the array is started no issues/error messages immediately:
But some times I continue to have the problem getting id and then sometimes it turns into the bellow btrfs error. This ahappens randomly after some time but reliable I can trigger it with an auto back up.

This is after the ca auto back up occurs I get the following error message:

ov 2 20:42:03 GLaDOS kernel: BTRFS error (device sdh1): bdev /dev/sdh1 errs: wr 0, rd 5525, flush 0, corrupt 0, gen 0
Nov 2 20:42:03 GLaDOS kernel: BTRFS error (device sdh1): bdev /dev/sdh1 errs: wr 0, rd 5526, flush 0, corrupt 0, gen 0
Nov 2 20:42:03 GLaDOS kernel: BTRFS error (device sdh1): bdev /dev/sdh1 errs: wr 0, rd 5527, flush 0, corrupt 0, gen 0
Nov 2 20:42:03 GLaDOS kernel: BTRFS error (device sdh1): bdev /dev/sdh1 errs: wr 0, rd 5528, flush 0, corrupt 0, gen 0
Nov 2 20:42:03 GLaDOS kernel: BTRFS error (device sdh1): bdev /dev/sdh1 errs: wr 0, rd 5529, flush 0, corrupt 0, gen 0
Nov 2 20:42:03 GLaDOS kernel: BTRFS error (device sdh1): bdev /dev/sdh1 errs: wr 0, rd 5530, flush 0, corrupt 0, gen 0
Nov 2 20:42:03 GLaDOS kernel: BTRFS error (device sdh1): error loading props for ino 4047063 (root 5): -5
Nov 2 20:42:04 GLaDOS kernel: BTRFS error (device sdh1): error loading props for ino 4047796 (root 5): -5

 

Attached is my diagnostic report if it helps.

 

If you guys/gals need anything else please let me know.

 

Thank you for your time!

 

Mark M.

glados-diagnostics-20211102-2043.zip

Edited November 3, 2021 by Markland

[JorgeB]

Nov  2 19:47:09 GLaDOS emhttpd: device /dev/sdf problem getting id
Nov  2 19:47:09 GLaDOS emhttpd: device /dev/sdc problem getting id
Nov  2 19:47:10 GLaDOS emhttpd: device /dev/sdg problem getting id
Nov  2 19:47:10 GLaDOS emhttpd: device /dev/sdd problem getting id

 

This are usually the result of having multipath to the enclosure enable, Unraid doesn't support that, use a single cable from the HBA to the enclosure.

 

Btrfs errors are the result of the device dropping offline, try using a different cables/slot or controller for the cache device.

 

 

[Markland]

On 11/3/2021 at 3:56 AM, JorgeB said:

Unraid doesn't support that, use a single cable from the HBA to the enclosure.

Awesome! I will try that when I get home.

 

On 11/3/2021 at 3:56 AM, JorgeB said:

cables/slot or controller for the cache device

When you say controller do you mean like the LSI controller or the one in the disk self?

 

Thank you for your help!

[JorgeB]

10 hours ago, Markland said:

like the LSI controller

this

[Markland]

On 11/3/2021 at 3:56 AM, JorgeB said:

Nov  2 19:47:09 GLaDOS emhttpd: device /dev/sdf problem getting id
Nov  2 19:47:09 GLaDOS emhttpd: device /dev/sdc problem getting id
Nov  2 19:47:10 GLaDOS emhttpd: device /dev/sdg problem getting id
Nov  2 19:47:10 GLaDOS emhttpd: device /dev/sdd problem getting id

 

This are usually the result of having multipath to the enclosure enable, Unraid doesn't support that, use a single cable from the HBA to the enclosure.

 

Btrfs errors are the result of the device dropping offline, try using a different cables/slot or controller for the cache device.

 

 

It has done the trick. Something of interest though is that If I have two cables hooked up to one of the controllers in the disk shelf I see all of the disks. If I hook it up to the other one I don't. And when I have both of the ports filled I don't get the "problem getting id" message and I dont get the BTRFS error any more.

 

I do have a new one though lol

 

Nov 7 00:02:44 GLaDOS kernel: blk_update_request: critical target error, dev sdc, sector 1950351263 op 0x3:(DISCARD) flags 0x800 phys_seg 1 prio class 0

 

I am curious if the two cables to one controller is causing that, or if its the controller as you had mentioned. Time for some more debugging and testing!

 

Thank you for your help!