LSI SAS Controller Questions

[elnevera]

Hi all,

 

I've been having some drive problems recently. 1 failed/not connected, 1 with pending sectors. New drives on route to swap in but I've noticed most of my drives in the array (7) have UDMA CRC error counts agains them.

 

My SAS controller (LSISAS2008) is showing in the diagnostics as;

 

May  3 17:44:16 Mystique kernel: mpt2sas_cm0: LSISAS2008: FWVersion(19.00.00.00), ChipRevision(0x03), BiosVersion(00.00.00.00)
May  3 17:44:16 Mystique kernel: mpt2sas_cm0: Protocol=(Initiator,Target), Capabilities=(TLR,EEDP,Snapshot Buffer,Diag Trace Buffer,Task Set Full,NCQ)
May  3 17:44:16 Mystique kernel: scsi host10: Fusion MPT SAS Host

 

Is it worth updating the formware? I haven't touched it since buying it and I can't even remember what brand it is.

Is there a way of telling what model it is without taking it out of the system?

I know I bought it from eBay but it is long gone from my purchase history.

 

If a firmware update will help, i'd like to make sure I get the correct model.

 

Thanks.

[elnevera]

D'oh! I checked my history on here and found the details of the card.

 

Description from the listing;

 

LSI SAS9211-8i 8-Port 6Gb/s SAS/SATA HBA Intergrated RAID PCI-e2 x8

"flashed to IT mode" works with Freenas/XigmaNas/Nas4Free etc

 

LSI is the brand?

 

[elnevera]

Is this what I need?

 

[JorgeB]

That firmware doesn't have any major issues AFAIK, but you should always upgrade to the latest, instructions here.

[elnevera]

43 minutes ago, JorgeB said:

That firmware doesn't have any major issues AFAIK, but you should always upgrade to the latest, instructions here.

 

Thank you, it was your instructions thread that I was looking at. Very useful!

 

I might as well try the update now I've learnt how to do it.

[elnevera]

Well, the firmware update went without a hitch, thanks very much @JorgeB

 

Once the array was back up, my 'not connected' disk has appeared in unassigned devices. I'm not sure what to make of that. Can someone please take a look at the attched SMART report and tell me if I should bin the drive or not?

 

ST4000VN008-2DR166_ZGY3DVHM-20220505-2128.txt

[ChatNoir]

23 minutes ago, elnevera said:

Can someone please take a look at the attched SMART report and tell me if I should bin the drive or not?

I am not seeing any red flags (but I am not an expert in SMART reports either ^^).

You should run an Extended SMART test on the drive to be sure. Disable spindown on the drive before or it will probably not go to the end.

 

Once it's finished, attach new diagnostics. (SMART reports are included with more information)

[elnevera]

On 5/5/2022 at 9:59 PM, ChatNoir said:

I am not seeing any red flags (but I am not an expert in SMART reports either ^^).

You should run an Extended SMART test on the drive to be sure. Disable spindown on the drive before or it will probably not go to the end.

 

Once it's finished, attach new diagnostics. (SMART reports are included with more information)

Sorry for the delay. I've done the above. Diagnostics attached.mystique-diagnostics-20220507-1731.zip

 

Thanks

[JorgeB]

Extended test pass so disk is OK, SMART attributes also look healthy.

[elnevera]

4 hours ago, JorgeB said:

Extended test pass so disk is OK, SMART attributes also look healthy.

Thank you very much!

[elnevera]

On 5/8/2022 at 9:54 AM, JorgeB said:

Extended test pass so disk is OK, SMART attributes also look healthy.

 

I am still getting UDMA CRC Errors on two drives. I will reseat the LSI card this weekend and check the cables to see if that helps.

[JorgeB]

Swap cables with different drives two see if the CRC errors follow the cables.

[sjoerd]

I would like to jump into this thread. Last Friday I received a new Kingston SA400 960G to use as cache pool for the domains (vms). I'm receiving udma crc errors instantly.

I got two LSI 92xx 8i controller in my rig - one is completely filled up and the 2nd has now 4 drivers (also 4 drives though the motherboard sata ports.

The the LSI controller are attached with 2 hff 87xx cable to the backplane on the enclosure. Since crc are not directly related to the drive (but communication between controller/ssd).

 

I would like to swap the drive (it's a cache pool ssd on a separate pool) to an other bay and see if the crc's disappear, but I'm not sure how to proceed. Do I need to shutdown the server, move to ssd to an other bay and start the server. One of the vm's is running on an unassigned drive. Can I e.g. stop the array, unmount de ssd, move it and then start the array again ?. The ssd is on the same cable as one of my parity-drives btw and that one is running fine. On the other cable of that same LSI controller are two unassigned disks connected, also fine

 

 

Edited October 5, 2022 by sjoerd

[JorgeB]

1 hour ago, sjoerd said:

Can I e.g. stop the array, unmount de ssd, move it and then start the array again ?.

That should work.

[sjoerd]

I swapped the drive to an other bay, even to the other cable of the 2nd controle - still same, meanwhile I contacted Kingston and the told me that this is not good and since the drive is brand new I need to get a new one from the shop I bought it.