Parity Check Error Count > 32k

[TwitchWink]

I've been using UnRaid for a while now.  I have a parity check happening monthly, usually returns passed with zero erros.  This last parity check returned 32963 errors.  So that seems like a lot more then normal.  😕

 

I'd appropriate any help looking through the diagnostics.  Not sure what to look for.

 

Thank you in advance.

aku-diagnostics-20200212-2135.zip

 

 

Looking into the Syslog it looks like I'm getting a lot of the following error in repeat.  Does this mean that drive is continually restarting?  How do I find out which drive is "ata14"?  Could this be causing my parity check errors?

 

Feb 10 07:54:09 AKU kernel: ata14.00: softreset failed (1st FIS failed)
Feb 10 07:54:12 AKU kernel: ata14.15: qc timeout (cmd 0xe4)
Feb 10 07:54:12 AKU kernel: ata14.00: failed to read SCR 0 (Emask=0x4)
Feb 10 07:54:12 AKU kernel: ata14.00: reset failed, giving up
Feb 10 07:54:12 AKU kernel: ata14.15: hard resetting link
Feb 10 07:54:15 AKU kernel: ata14.15: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
Feb 10 07:54:15 AKU kernel: ata14.00: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
Feb 10 07:54:16 AKU kernel: ata14.01: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
Feb 10 07:54:16 AKU kernel: ata14.02: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
Feb 10 07:54:16 AKU kernel: ata14.03: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
Feb 10 07:54:17 AKU kernel: ata14.04: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
Feb 10 07:54:17 AKU kernel: ata14.00: configured for UDMA/33
Feb 10 07:54:17 AKU kernel: ata14.01: configured for UDMA/133
Feb 10 07:54:17 AKU kernel: ata14.02: configured for UDMA/133
Feb 10 07:54:17 AKU kernel: ata14.03: configured for UDMA/133
Feb 10 07:54:17 AKU kernel: ata14.04: configured for UDMA/133
Feb 10 07:54:17 AKU kernel: ata14: EH complete

 

Edited February 13, 2020 by TwitchWink

[JorgeB]

3 hours ago, TwitchWink said:

which drive is "ata14"?

It's not just ata14, it's ata14.00 to ata14.04, the five disks on the port multiplier, SATA port multipliers are not recommended, though unlikely it's causing the sync errors, but if nothing else it will be causing performance issues.

 

Also because of the log spam caused by that there's nothing else to see, can't even see if it was a correcting check, first step would be to run a correcting check and then a non correcting one to see if the errors keep happening, ideally after getting rid of the port multiplier.

[TwitchWink]

I have two Syba SD-PEX40104 PCI-E Cards driving my UnRaid.  lspci quote below for these controllers.  The boards have 8 SATA Ports each and each port only has one hard drive per port.

 

These controllers use Dual Marvell 88SE9215 + 88sm9705 chipsets, spec sheet from Marvel.com. 

 

Looking around online I found this FAQ about port multipliers at sata-io.org.  My UnRaid system looks like the top drawing, where each hard drive is directly connected to a SATA port on a Syba Card via a single SATA cable. 

 

Does the Syba controller act as a Port Multiplier?

 

Quote

36:00.0 SATA controller [0106]: Marvell Technology Group Ltd. Device [1b4b:9215] (rev 11)
    Subsystem: Marvell Technology Group Ltd. Device [1b4b:9215]
    Kernel driver in use: ahci
    Kernel modules: ahci
37:00.0 SATA controller [0106]: Marvell Technology Group Ltd. Device [1b4b:9215] (rev 11)
    Subsystem: Marvell Technology Group Ltd. Device [1b4b:9215]
    Kernel driver in use: ahci
    Kernel modules: ahci

 

[JorgeB]

4 minutes ago, TwitchWink said:

Does the Syba controller act as a Port Multiplier?

It includes a port multiplier, the Marvell 9215 is a 4 port controller, one of them connects to the port multiplier, so then you have 3 + 5 ports, those 5 are sharing a single SATA port.