disk in error state. first time here, do i need to replace?

[giantrobot]

Hi there,

 

I've got a server that's been running fine for about 8 weeks with no problems. I accidentally powered it off three days ago, and finished a successful parity check upon restart (with errors detected). That being resolved, I put in two 8TB WD Red drives into the array and started preclearing them this morning, only to come down with errors on one of my 10TB drives, taking it offline.

 

Is there anything I can do to not replace the 10TB drive, or do I at the very least need to preclear it again for use? If I do need to replace it, once the 8TB drives come online, is there a way to move stuff around so I can evacuate enough data off of that 10TB drive and replace it with one of the 8TB ones?

 

Running:

Version: 6.7.2 Plus

ITX build 
Intel Corporation DQ77KB

Intel® Core™ i5-3570S CPU @ 3.10GHz

8 GiB DDR3

Parity WDC_WD100EFAX-68LHPN0_JEKUVZSZ - 10 TB 

Disk 1 WDC_WD100EFAX-68LHPN0_JEKW3EKZ - 10 TB

Disk 2 WDC_WD40EFRX-68WT0N0_WD-WCC4E1799698 - 4 TB

(offline) Disk 3 WDC_WD100EMAZ-00WJTA0_JEHMZHWM - 10 TB

Disk 4 WDC_WD40EFRX-68WT0N0_WD-WCC4E2069589 

Disk 5 WDC_WD40EFRX-68WT0N0_WD-WCC4E6UR35SY 

 

Dockers:

Plex, Medusa, Tautulli, Resilio Sync, Krusader

Plugins:

Preclear, Rclone

voltron-diagnostics-20191014-0354.zip

[trurl]

On mobile now so can't look at Diagnostics. Please wait for advice. 

[JorgeB]

Disk dropped offline so there's no SMART, and although I'm also on phone and not so easy to check diags looks like a controller problem, you're using an Asmedia controller with a port multiplier and those are known to have issues and are not recommended, suggest replacing it with an LSI HBA.

[JorgeB]

Forgot to mention that you'll need to reboot/power cycle the server for the dropped disk to come online, then post new diags so we can check SMART but most likely it's fine.

[giantrobot]

3 hours ago, johnnie.black said:

Forgot to mention that you'll need to reboot/power cycle the server for the dropped disk to come online, then post new diags so we can check SMART but most likely it's fine.

is it worth letting the preclear on the 8TB drive finish first, or should i restart now so i can get good SMART reports?

[giantrobot]

3 hours ago, johnnie.black said:

Disk dropped offline so there's no SMART, and although I'm also on phone and not so easy to check diags looks like a controller problem, you're using an Asmedia controller with a port multiplier and those are known to have issues and are not recommended, suggest replacing it with an LSI HBA.

ugh, what? rats, i thought all i had to do was to avoid Marvell controllers. does LSI even make a PCIe 4x controller? alll the ones i see are 8x and won't fit on a DQ77KB.

[JorgeB]

2 minutes ago, giantrobot said:

does LSI even make a PCIe 4x controller?

Not SAS2/SAS3, and older ones can't handle >2TB disks.

 

Looking now at the diags again, disk4 also dropped offline, not surprisingly it's using the same controller.

[JorgeB]

Also two port Asmesdia controllers are fine, 6 and 10 port Asmedia controllers use port multipliers and those are nothing but trouble.

[giantrobot]

1 hour ago, johnnie.black said:

Also two port Asmesdia controllers are fine, 6 and 10 port Asmedia controllers use port multipliers and those are nothing but trouble.

anyone know of an option to get 5-6 SATA ports out of a PCIe 4x slot? my case has 8 drive bays; with additional SSDs for cache that's 9-10 devices, so besides the on-board 4 SATA ports i would like to get 5-6 so I can use all the bays...

[JorgeB]

Can't think of any good option, you could cut/dremel the PCIe slot to fit an x8 LSI.

 

 

 

 

 

Edited October 14, 2019 by johnnie.black

[giantrobot]

14 minutes ago, johnnie.black said:

Can't think of any good option, you could cut/dremel the PCIe slot to fit an x8 LSI.

 

 

 

 

 

omg yikes.

 

could i not just try one of these: SuperMicro AOC-SASLP-MV8 8-port PCIe x4 SATA II Marvell 6480 listed on the https://wiki.unraid.net/Hardware_Compatibility#PCI_SATA_Controllers section?

 

[JorgeB]

Just now, giantrobot said:

could i not just try one of these: SuperMicro AOC-SASLP-MV8 8-port PCIe x4 SATA II Marvell 6480 listed on the https://wiki.unraid.net/Hardware_Compatibility#PCI_SATA_Controllers section?

Those are not recommended at all for v6, they worked OK with v4/v5.

[giantrobot]

ok, i finished my one good preclear, removed the extra two drives, and restarted the machine. here are the latest diagnostics. all smart reports claim PASSED -- maybe everything was just the controller being overloaded and all is good? if so, how do i put it back into operation? do i follow the procedure to unassign/reassign it to the array and rebuild onto itself?

voltron-diagnostics-20191015-0038.zip

[JorgeB]

6 hours ago, giantrobot said:

maybe everything was just the controller being overloaded and all is good?

With those controllers it will happen with normal use, and a good controller can't really be overloaded, you should be able to use all ports at once.

 

6 hours ago, giantrobot said:

if so, how do i put it back into operation? do i follow the procedure to unassign/reassign it to the array and rebuild onto itself?

You can, but if you don't replace that controller it will likely happen again sooner or later.