How to find the disk with errors from a log with only SATA errors ?

[Singman]

Hi everyone,

I have a disk that will fail soon and make Proxmox server crash randomly.
On the console, the only error I could find is :



I found that ATA6.00 is /dev/sdc and I could confirm it :

root@pve2:~# smartctl -a /dev/sdc
smartctl 7.3 2022-02-28 r5338 [x86_64-linux-6.8.12-10-pve] (local build)
Copyright (C) 2002-22, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Model Family:     Crucial/Micron Client SSDs
Device Model:     CT1000MX500SSD1
Serial Number:    2212E61F2F8B
LU WWN Device Id: 5 00a075 1e61f2f8b
[...]
199 UDMA_CRC_Error_Count    0x0032   100   100   000    Old_age   Always       -       13

My question is simple : how to find the /dev/sdc disk when I have 4x SSD disks of the same type in the server ?
Disassemble them one by one and compare to the serial number smartctl return ?

 

[Kingneutron]

You can track which disks are in which slots by using a spreadsheet. Can even take a picture of the drive label and put it in a cell and track warranty expiration dates.

https://github.com/kneutron/largefiles

https://github.com/kneutron/ansitest/blob/master/drivemap.sh
.

Yes, this requires some downtime but you'll need to do it for server maint and disk replacement anyway.

 

[alexskysilk]

how to find the /dev/sdc disk when I have 4x SSD disks of the same type in the server ?

lsscsi --verbose

you're looking for bus id 6.00

but with only 4 drives just run a full smart test on ALL of them.

 

[gfngfn256]

Adding to Kingneutron's documentation advice (& nice spreadsheet!): Also place a label/sticker containing the drive s/n in a VISIBLE & ACCESSIBLE place (even while installed) on the drive/bay for easy identification. This is a must for any server & especially for hot-plugged drives.

 

[celtar]

Hi everyone,

I have a disk that will fail soon and make Proxmox server crash randomly.
On the console, the only error I could find is :

View attachment 86001

I found that ATA6.00 is /dev/sdc and I could confirm it :

root@pve2:~# smartctl -a /dev/sdc
smartctl 7.3 2022-02-28 r5338 [x86_64-linux-6.8.12-10-pve] (local build)
Copyright (C) 2002-22, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Model Family:     Crucial/Micron Client SSDs
Device Model:     CT1000MX500SSD1
Serial Number:    2212E61F2F8B
LU WWN Device Id: 5 00a075 1e61f2f8b
[...]
199 UDMA_CRC_Error_Count    0x0032   100   100   000    Old_age   Always       -       13

My question is simple : how to find the /dev/sdc disk when I have 4x SSD disks of the same type in the server ?
Disassemble them one by one and compare to the serial number smartctl return ?

Hi,

here is an tutorial: https://forum.proxmox.com/threads/zfs-austauschen-einer-disk.135765/#post-761665

I think Post 11 is the right one.

"apt install ledmon -y "
"ledctl locate=/dev/sdc" # turns light on Disk

Just my 2cent
Thanks
John

 

[gfngfn256]

Just my 2cent

Excellent advice - but will only work if drive/bay has independent light for each disk activity.

 

[Singman]

Ya, don't work with SSD
But i'm lucky, 1st disj unmounted and that was the right one !