Parity Sync Issue

[swells]

Hi All,

 

Looking for some help with an issue. I searched around here a bit and couldn't find anything specific to what I am seeing, but if there is something out there, please forgive me and point me in the right direction please.

 

I am having an issue that has occurred since the beginning of the year. I have been having an issue with parity sync and have been getting 5 sync errors every other month or so, then I correct them usually after I confirm they are there, and then wont see them for a month or so and then they come back. And they are always in the same sectors. I started tracking the sector a few months ago and copied them off into a separate text file to keep. Below are the details. Current state of my server is, I just within the past two weeks transferred all hardware from a Antec 900 with three 5in3 cages to a norco 4224. So all cables, back planes, caddies, etc are all different since the last time I saw the sync errors. If it was memory I would think it would be at different sectors right? Could it be bad sectors on a single drive? What do I need to do? Long SMART test on each drive? All the SATA cards stayed the same, I have a Dell PERC H310 and a single 2 port card coupled with my 6x on board ports currently in use. When I moved to the norco chassis I added an additional H310 but it is currently not in use. 

 

As you can see here, I will get the sync errors, run again with correct again, they get corrected and then i get a month or more of no errors and then they return. 

2018-09-21, 14:34:05    19 hr, 35 min, 41 sec    113.4 MB/s    OK    0
2018-09-18, 08:56:43    20 hr, 49 min, 55 sec    106.7 MB/s    OK    5
2018-08-02, 01:59:33    23 hr, 59 min, 32 sec    92.6 MB/s    OK    0
2018-07-01, 22:05:41    20 hr, 5 min, 40 sec    110.6 MB/s    OK    5    
2018-06-01, 20:42:18    18 hr, 42 min, 17 sec    118.8 MB/s    OK    0    
2018-05-21, 15:50:31    18 hr, 29 min, 50 sec    120.2 MB/s    OK    5    
2018-05-19, 04:26:08    18 hr, 37 min, 32 sec    119.3 MB/s    OK    5    
2018-05-18, 01:45:21    18 hr, 44 min, 55 sec    118.6 MB/s    OK    0    
2018-05-01, 19:08:03    21 hr, 34 min, 27 sec    103.0 MB/s    OK    0    
2018-04-29, 15:57:07    21 hr, 7 min, 35 sec    105.2 MB/s    OK    0    
2018-04-28, 16:28:48    20 hr, 49 min, 35 sec    106.7 MB/s    OK    0    
2018-04-27, 18:42:46    20 hr, 42 min, 24 sec    107.3 MB/s    OK    5    
2018-04-01, 22:57:58    20 hr, 57 min, 57 sec    106.0 MB/s    OK    0    
2018-03-03, 03:38:01    20 hr, 55 min, 32 sec    106.2 MB/s    OK    5    
2018-03-01, 23:02:53    21 hr, 2 min, 52 sec    105.6 MB/s    OK    5    
2018-02-03, 02:48:41    20 hr, 43 min, 30 sec    107.2 MB/s    OK    5
2018-02-01, 22:46:09    20 hr, 46 min, 8 sec    107.0 MB/s    OK    5
2018-01-01, 23:16:40    21 hr, 16 min, 39 sec    104.5 MB/s    OK    5

 

In July I started tracking the sector of the errors. 

Jul  1 05:30:30 Tower kernel: md: recovery thread: P incorrect, sector=2743151176
Jul  1 05:30:30 Tower kernel: md: recovery thread: P incorrect, sector=2743151184
Jul  1 05:30:30 Tower kernel: md: recovery thread: P incorrect, sector=2743151192
Jul  1 05:30:30 Tower kernel: md: recovery thread: P incorrect, sector=2743151200
Jul  1 05:30:30 Tower kernel: md: recovery thread: P incorrect, sector=2743151208

 

Jul 2 08:55:30 Tower kernel: md: recovery thread: P incorrect, sector=2743151176
Jul 2 08:55:30 Tower kernel: md: recovery thread: P incorrect, sector=2743151184
Jul 2 08:55:30 Tower kernel: md: recovery thread: P incorrect, sector=2743151192
Jul 2 08:55:30 Tower kernel: md: recovery thread: P incorrect, sector=2743151200
Jul 2 08:55:30 Tower kernel: md: recovery thread: P incorrect, sector=2743151208

 

Oct  1 04:46:49 Tower kernel: md: recovery thread: P incorrect, sector=2743151176
Oct  1 04:46:49 Tower kernel: md: recovery thread: P incorrect, sector=2743151184
Oct  1 04:46:49 Tower kernel: md: recovery thread: P incorrect, sector=2743151192
Oct  1 04:46:49 Tower kernel: md: recovery thread: P incorrect, sector=2743151200
Oct  1 04:46:49 Tower kernel: md: recovery thread: P incorrect, sector=2743151208

 

Any ideas? Any help here would be appreciated. 

[JorgeB]

32 minutes ago, swells said:

and a single 2 port card

Is this a Marvell controller? If you don't know post your diagnostics: Tools -> Diagnostics

[swells]

Looks like it is. This is the card. One of the reviews states the card has a Marvell 88SE9128.
https://www.amazon.com/gp/product/B003GS8VA4/ref=oh_aui_search_detailpage?ie=UTF8&psc=1

 

And this is from my diagnostics. Marvell no bueno?

04:00.0 SATA controller [0106]: Marvell Technology Group Ltd. 88SE9128 PCIe SATA 6 Gb/s RAID controller [1b4b:9128] (rev 20)
    Subsystem: Marvell Technology Group Ltd. 88SE9128 PCIe SATA 6 Gb/s RAID controller [1b4b:9128]
    Kernel driver in use: ahci
    Kernel modules: ahci

[JorgeB]

5 minutes ago, swells said:

Marvell no bueno?

No, and one of the known issues that can sometimes happen is getting 5 parity sync errors every time after a power cycle/reboot.

[swells]

Well that explains that. I can move the drive(s) that are connected to it to free slots on the H310. Is there a 2 port SATA3 card that is recommended to replace it?

[weirdcrap]

I don't know of any specifically to recommend but the Wiki has a section that should get you started in your search: https://wiki.unraid.net/Hardware_Compatibility#PCI_SATA_Controllers

Edited October 1, 2018 by weirdcrap

[JorgeB]

2 hours ago, swells said:

Is there a 2 port SATA3 card that is recommended to replace it?

Use an Asmedia based controller, e.g.

https://www.amazon.com/IO-Crest-2-port-Controller-SY-PEX40039/dp/B005B0A6ZS/ref=sr_1_1?ie=UTF8&qid=1538432850&sr=8-1&keywords=SYBA+SY-PEX40039

[swells]

 Thanks for your help!