Onboard SATA ICH8R took a dump or board / processor silently failed - story time

[cpetro45]

I've been running Unraid for 10 years.  I have a 6 disk array that at one time was 9 or 10.  All has been well.  As you can imagine drive failures here and there, etc over 10 years.  I updated disks and consolidated a few years ago.  I have 8TB parity and 5 Data disks 2 - 4 GB each.  I never updated the H/W.  This thing has been running on Core 2 Duo E6600 or something with 2 GB of ram...ancient Gigabyte board and an 8800GTX , still no issues besides regular disk maintenance...til now.

 

Log in to WebUI the other day and see a disabled disk...I'm pretty annoyed as nothing has changed and don't even write to this that much.  I didn't grab the logs.  The SMART report on disabled disk says UDMA CRC error count.  I read all about that all over this forum.  In the meantime, I think I stopped the array and shutdown to replace SATA cables.  I ordered new SATA cables from Amazon.  I replaced all of them (they are old).  I didn't care about the sata plug / disk order cause unraid 6 no need to worry.  I think I was concerned so I hooked up a monitor as well to check booting.  Well, I thought it got stuck on loading bzroot since it was taking forever (more than 3 - 4 min).  I got sidetracked and ended replacing the usb drive (14 years old) thinking that was bad (it wasn't).  I had to use the HP Media preparation since this MOBO is real picky.  Did all that BS (figuring that out again that it wouldn't boot unless specially formatted with a 15 year old tool), finally got it booting from new USB thumb drive.  BTW, it was taking like 10 minutes to boot (first hint, since it never usually took that long)...remember this thing is still booting up.  I didnt think it was booting and thought usb drive was bad.  I used the new UNRAID 6 usb drive prep first, obviously, but it definitely needed that HP tool (thanks to this forum...again... I was able to find that info ...again and re-download it).

 

 

Next, the forums say, oh, just check the file system of the disabled disk, and it's probably that.  I started array in maintenance mode (after 10 minute boot or so), ran the file system check, nothing really came up on -n option.  I figured, what the hell, run it without -n.  Next thing you know, I get a UDMA CRC error count warning on PARITY disk, during the File system check of disabled disk.  I'm like , oh man, parity better not get disabled.  There were I think 21 writes in the Parity disk column and 20 errors.  I started getting real nervous.  Somehow the retries must have been successful (I have that syslog)  I cancelled the file system check on disabled disk.  I had screenshots from config before and after I moved the disks around.  It wasn't the same SATA port that threw the UDMA CRC error count on data disk and parity disk.  At this point, I knew something on the board was toast. 

 

Luckily, I have another system (literally only 2 years newer) but with 8GB ram and Core 2 Xtreme 3000 quad core (1200$ processor back in it's day , bought on ebay for like 250 or something like 8 years ago).  Anyway, huge upgrade compared to the 2GB and Core 2 Duo LOL.  We're talking DDR 2 here. 

 

So, conveniently enough, the "new" computer had a removable MOBO tray, so right now, I'm running the array off the new board, rebuilding the failed disk on top of itself.  Something went real bad on in the old CPU.  Added some pics....hope this was a fun blast from the past for some, and also BACK UP YOUR DATA because you never know...

 

Ultimately, I do feel like a few bits may be off (I hope not).  I definitely going to backup all the rest of my important items once array is back up completely. 

 

 

Edited February 26, 2022 by cpetro45