Unraid Parity Sync slow

[jomack16]

As you can see from the image, The parity sync is going to take quite a while. I suspect that either a drive or a connection to a drive is causing the slowness, but I am unsure how to check that. I have attached the diagnostics for anyone that is kind enough to check them out for me.

note* for anyone who cant see the image, or if it doesn't load for some reason, here are the numbers:

Total Size: 12 TB
Elapsed time: 11 minutes
Current position: 44.9MB (0.0%)
Estimated speed: 47.6KB/sec
Estimated finish: 2903 days, 17 hours, 40 minutes

jej-unraid-diagnostics-20210120-1057.zip

[trurl]

Problems with multiple disks, 8,9,10,11,12,13. 12 and 13 aren't even showing up in SMART. Probably controller or power issue.

 

You have a Marvell controller. Are the problem disks on that controller?

[jomack16]

Yes, it is an 8 port controller (https://www.sybausa.com/index.php?route=product/product&product_id=160)
 

[JorgeB]

That controller is not recommended, you also have several disks connected with USB, that's even worse, and appears to be the main issue currently.

[jomack16]

Thanks for looking! I am soon to change the USB connection configuration, but I did not know that the controller was not recommended.
The disks that are connected via USB are not currently in the array, would they still be causing the slowness on parity Sync?

Also I am considering replacing that controller with this one instead: https://www.ebay.com/itm/ASR-72405-Adaptec-24-Port-1GB-Cache-6Gb-s-SAS-SATA-PCIe-3-RAID-Controller-Card/183570011464?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649
 

Thoughts?

[JorgeB]

14 minutes ago, jomack16 said:

The disks that are connected via USB are not currently in the array

When is currently? I looked at the beginning of the syslog in the diags, disks 1, 8, 9, 10, 11, 12 and 13 were connected with USB, then they dropped:

 

Jan 15 13:01:40 JEJ-Unraid kernel: usb 2-4: Device not responding to setup address.
### [PREVIOUS LINE REPEATED 1 TIMES] ###
Jan 15 13:01:41 JEJ-Unraid kernel: usb 2-4: device not accepting address 2, error -71
Jan 15 13:01:45 JEJ-Unraid kernel: usb usb2-port4: Cannot enable. Maybe the USB cable is bad?
### [PREVIOUS LINE REPEATED 2 TIMES] ###
Jan 15 13:01:53 JEJ-Unraid kernel: usb 2-4: USB disconnect, device number 2

 

Then the disks errors:

 

Jan 15 13:01:53 JEJ-Unraid kernel: md: disk1 write error, sector=4294967280
Jan 15 13:01:53 JEJ-Unraid kernel: md: recovery thread: exit status: -4
Jan 15 13:01:53 JEJ-Unraid kernel: md: disk8 read error, sector=8633241376
Jan 15 13:01:53 JEJ-Unraid kernel: md: disk9 read error, sector=8633241376
Jan 15 13:01:53 JEJ-Unraid kernel: md: disk10 read error, sector=8633241376
Jan 15 13:01:53 JEJ-Unraid kernel: md: disk11 read error, sector=8633241376
Jan 15 13:01:53 JEJ-Unraid kernel: md: disk12 read error, sector=8633241376
Jan 15 13:01:53 JEJ-Unraid kernel: md: disk13 read error, sector=8633241376

 

Array did change after that, but that part is not covered in the diags due to log spam so we can't really see the problem, reboot and post new diags

[JorgeB]

22 minutes ago, jomack16 said:

Also I am considering replacing that controller with this one instead: https://www.ebay.com/itm/ASR-72405-Adaptec-24-Port-1GB-Cache-6Gb-s-SAS-SATA-PCIe-3-RAID-Controller-Card/183570011464?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

 

I'm more familiar with LSI controllers, but IIRC others users are using that one, still best to search the forum to confirm.

[jomack16]

I was finally able to reboot and get new diagnostic logs. Attached here. Curious if it shows anything other than the USB drives as an issue.

jej-unraid-diagnostics-20210120-1512.zip

[danejaho]

I would highly recommend staying with the tried and true LSI/Dell SAS backplane cards in IT mode. This is by far the most common HBA in the Unraid community and they prove to be extremely reliable.

 

Most HBAs only come with two SAS ports, but if you watch this you can see how to add up to 120 HDDs using a single HBA card.

 

[trurl]

Did you still have any USB disks connected when you took those diagnostics? 

[jomack16]

I did. I am not physically at the server location, so I couldnt unplug the cable. Is that still mucking up the diagnostics?

[trurl]

Something is mucking it up. Are the USB connected disks assigned?

[jomack16]

The USB disks are not assigned, no.

[trurl]

Port multiplier ata10 having a lot of problems in syslog. Don't know if that is the Marvell or not. Some people are able to have success with Marvell by disabling IOMMU (vt-d for intel) in the BIOS. Do you have any VMs that need that?

[JorgeB]

 

7 hours ago, trurl said:

is the Marvell

Yes, the problem now is the Marvell controller with port multipliers.

[jomack16]

Awesome! Thanks so much for the help!
I'll be replacing that controller as soon as possible with something from here: https://wiki.unraid.net/Hardware_Compatibility#PCI_SATA_Controllers

Also, just in case someone comes across this in the future, This is the card that is causing the trouble: 8 Port SATA III PCE-e 2.0 x2 Card - SI-PEX40071

[JorgeB]

Recommend you get one the recommended LSI controllers, any LSI with a SAS2008/2308/3008/3408 chipset in IT mode, e.g., 9201-8i, 9211-8i, 9207-8i, 9300-8i, 9400-8i, etc and clones, like the Dell H200/H310 and IBM M1015, these latter ones need to be crossflashed.