Any insight to this problem below is much appreciated.

 

Began the process of rebuilding parity after removing a data drive earlier today. (Removed drive 1 after moving all data off of it). Did a "New Config" and started the array to do parity rebuild.

Almost immediately, began seeing the following error in the syslog.

 

Apr 3 08:12:44 TOWER ntfs-3g[30147]: ntfs_attr_pread_i: ntfs_pread failed: Input/output error

Apr 3 08:12:44 TOWER ntfs-3g[30147]: Failed to read index block: Input/output error

Apr 3 08:12:44 TOWER kernel: Buffer I/O error on dev sdl1, logical block 61047301, async page read

 

I am baffled because there is no device sdl1 in the array, and as far as I can  tell that device wasn't there even before the new config.

I copied all device assignments down prior to doing the New config, and no sdl1.

As of now there is still no device sdl1, so how can there be a failure on a device that does not exist.

 

In addition, the parity rebuild is continuing despite the errors, and appears to be running fine and scheduled to complete in about 50 minutes. This would put entire rebuild time around 10 hours for 3 x 4TB and 1 x 2TB array, well above the 100mb/sec mark.

 

I am attaching diagnostics which show the errors and sequence of events.

 

Lastly, I am running 6.6.7.

 

Thanks in advance for any help.

tower-diagnostics-20190403-1635.zip

Update: The Parity Rebuild process completed on time with no errors. Stopped all services and rebooted. All is up and running, and no errors showing in the log yet. I am monitoring and will post update after some time has passed.

 

I have no idea how errors were happening on a device that was not in the array, so any insight is appreciated.

sdl1 was an NTFS unassigned device, so nothing to to with the array:

Apr  2 11:39:20 TOWER unassigned.devices: Successfully mounted '/dev/sdl1' on '/mnt/disks/Hitachi_HUA723020ALA640_MK0271YGJ7RVHA'.

 

(Solved) Thanks for clearing up the confusion