Parity rebuild very slow

[Housemann]

Hello everybody,

 

i have a problem that my parity sync is too slow. 
I created an array with 6x 3TB and 2x 1TB. Then i move some Files to the array that fills up onto 3 TB. 
I used for my movies allocation method: fill up. Later i have inserted an 6 TB parity drive but the rebuild process is very slow. 
Unraid shows me more then 500 days for parity rebuild. The rebuildrate are 200-300 kb/s.
The array read speed are by 100-140 mb/s. Write speed is near to the same on the array.

Unraid shows me at two of the discs Errors (one with 159 and one with 81536 Erros (i must replace this discs).ö

I dont know if the parity build is slow because the errors on the two discs?
If i copy data from disks with errors on my NAS, it is very slow. The rate are 1,4 MB/s. 
Is this the Problem, that these discs the slowest on the array and that is the speed the parity discs work?

 

I dont know, if my config is wrong or i have to many components for the PCIe Lanes or because the errors on the discs?

My System have the folloing components.

CPU --> Intel Xeon E-2146G
RAM --> 4x Crucial 16GB DDR4-2666 UDIMM - CT16G4DFD8266

Harddrives
    --> 6x WD RED 3TB Array
    --> 2x WD RED 1TB Array
    --> 1x WD RED 6TB Parity

Asus Pro WS C246-ACE
    2 x PCIe 3.0/2.0 x16 (x16 or dual x8)    
        --> HP H220 6Gbps SAS PCI-E 3.0 HBA LSI 9207-8i P20 IT Mode
        --> IBM ServeRAID 16-Port 6Gbps SAS-2 SATA Expansion Adapter 46M0997 Firmware 634A
    1 x PCIe 3.0/2.0 x16 (max at x2 mode)
        --> HP - 593717-B21 - NC523SFP 10Gb Dual Port Server Adapter
    2 x PCIe 3.0/2.0 x1
        --> Nothing
    
    1 x U.2 connector 
        --> 4x Samsung SSD evo 250gb for Cache Drive -- Unraid Raid10
        
    1 x M.2_1 socket 3, , with M key, type 2242/2260/2280/22110 storage devices support (PCIE 3.0 x 4 mode)
        --> KINGSTON_SA1000M8480G_50026B7682D6B9C6 M2SSD 480gb
    1 x M.2_2 socket 3, with M key, type 2242/2260/2280/22110 storage devices support (SATA & PCIE 3.0 x 2 mode)
        --> Nothing

    
Kind regards
Oliver

 

[trurl]

Go to Tools-diagnostics and attach the complete Diagnostics ZIP file to your NEXT post 

[Housemann]

Here are the File.

skynet-diagnostics-20200609-1524.zip

[itimpi]

Both desk3 an disk6 are getting continual read errors so not much point in trying to continue the parity sync.   Both drives also show pending sectors in their SMART reports which indicate some sectors that cannot currently be read reliably.

 

I would suggest that the following steps are a good idea:

• Check the power/SATA cabling to the drives as this is the commonest cause of disk problems
• Run the extended SMART test on both drives to check them out.

After doing that you might want to post new diagnostics so we can see what the SMART tests reported.

 

[Housemann]

The power and SFF-8087 cables are permanently connected on the SAS Card and backplane.
The six backplanes have seperatly power an SFF ports.

The SMART Test on disc 3 results "Errors occurred - Check SMART report". Also disc 6.

Here are the results...

 

 

skynet-diagnostics-20200610-0450.zip

[itimpi]

According to the SMART reports both disk3 and disk6 are faulty and need replacing.

[Housemann]

Thanks for reply.

My colleague has two, almost new, wd red 4tb drives i can have. I replace it an give a feedback. 

[sota]

hopefully they're not SMR reds...

[JorgeB]

SMR disks usually work fine with Unraid since it's not RAID.