Parity Drive Keeps Error'ing

[darkcyde]

Hi

 

Bit of background first - I've just got a HP DL180 G6 and I'm using an array of 8 2TB SAS 7.2k drives, with one as Parity.  I'm using the included P410 array card assigning each drive as an individual Logical and passing these to unRaid....I know as a drawback here is I don't get any SMART feedback from the drives, as well as other drawbacks.

 

When I first setup the array, I just allowed unRaid to clear each drive as I wasn't aware of the preclear script....everything checked out, and was running fine for about 5 days until the Parity drive developed about 111 errors and unRaid took it offline.  I figured it was just a duff drive.

 

So I took an unused drive from the end of the array and replaced this as Parity, this time doing a full Preclear with the plugin/script, which took about 18 hours, and completed fine.  I added this drive into the array and assigned it as Parity, restarted the array and it commenced building the Parity.

 

It got about an hour and a half in before this drive started throwing up errors and unRaid took it offline again.

 

I'm currently preclearing the old Parity drive to see if it clears properly.

 

Should preclearing show up any drive errors?  I'm using the full script so pre-read and post-read, which was all fine.

 

Why is this only happening with the Parity drive?  Do I have a duff socket in the backplane?  I know without any SMART data we can only guess, but I need to understand how unRaid is picking up these errors without SMART capability....is it just purely a read/write test that is failing, if so why did preclear complete fine, or is this something else?

 

In the logs I'm getting this....

 

Nov 4 11:56:03 ******* kernel: print_req_error: critical medium error, dev sdc, sector 481156720

Nov 4 11:56:03 ****** kernel: md: disk0 write error, sector=481156656

and then a whole load of write errors and different sector codes

 

I'm just not clear in my head why it would clear, but then throw up these write errors.

 

Would I be better off changing to a H220 card which would allow me to have SMART capabilities, but I've read some horror stories of people trying to get this server and card combo working properly.  Would the H220 card just plug in and be a directly replacement for the P410 on my 12 bay backplane or would I need new cables, and/or an expander too?

 

Thanks in advance for your help!

 

Leo

[JorgeB]

Please post your diagnostics: Tools -> Diagnostics, though because of the RAID controller they might not be much help.

[darkcyde]

Here they are, but there is zero SMART data... I've tried running smartctl but it returns nothing. 

 

I've bitten the bullet and ordered a refurbed H220, and started backing up all my data in prep for a complete array reset. 

 

I checked the specs for both the P410 and H220 cards, and they use the same connectors....the H220 is a Gen 8 card, but people are reporting it working in the Gen 6 servers, so fingers crossed

diagnostics-20181104-1719.zip

[JorgeB]

Yeah, without the SMART data not much help, if issues continue with the H220 then post new diags.

[darkcyde]

Ok, so I got a H220 and plugged it in, and now the disks are all handled by unRaid, and I can see SMART data, so fingers crossed the Parity holds this time.

 

I had to flash the H220 to get it to accept newer drives like the SSD, so I used the instructions from here, just in case anyone wants to do the same thing!