ata errors when under stress?

megamind6155 · 2022-06-17 10:02:53

Hello,

I have a hdd setup which has my minecraft instances in it.
I run my system of of a ssd.

when I run minecraft in a stressful way (this maybe applicable to other applications too)

I get harddrive errors.
the type of errors I get in dmesg:
http://ix.io/40lk

I did do a harddrive test using smartctl
a short one and a long one both were fine
here is smart log:
http://ix.io/40lm

if I run minecraft for too long I suspect data corruption also happens on the drive. It has happened a few times before too.
I have to fix the filesystem using fsck but I've had some data loss.

curiously it only happens when under some sort of load?
Also the drive is pretty new like It has been one year since I bought it.

Help would be appreciated.

seth · 2022-06-17 13:48:46

183 Runtime_Bad_Block       0x0032   099   099   000    Old_age   Always       -       1

Keep an eye on that value. Other than that the drive looks ok.

Bus error could be badly seated/cable or - given the sympton description - underpowered/overheated
Why is there "libata.noacpi=1"?

megamind6155 · 2022-06-18 03:23:23

There was another ata error but the value didn't change.

I put libata.noacpi=1 to get rid of an error.
I thought the error was causing the harddrive problem so I looked it up.
I found you can add this parameter and it goes away. I didn't lose any functionality so I thought it was fine.
Here is the error I used to get:
http://ix.io/40ox

It maybe a underpowering problem/overheating problem. I can't tell.
I've tried to reset the cable. I've put it in different sata slots. It doesn't help.

frostschutz · 2022-06-18 06:37:29

Is your partitioning okay? It seems to complain about unaligned writes among other things.

The bus speed resets to 3.0G (SATA 2), you could force ata5 to that speed with libata.force=5:3.0G

I've tried to reset the cable. I've put it in different sata slots. It doesn't help.

Did you replace the cable altogether?

Last edited by frostschutz (2022-06-18 06:38:19)

megamind6155 · 2022-06-19 10:25:47

I dunno how to check partitioning
here is lsblk:

NAME   FSTYPE FSVER LABEL UUID                                 FSAVAIL FSUSE% MOUNTPOINTS
loop0
└─swap swap   1     swap  e72d2729-20ba-45a0-93ed-c418aa52b5f2                [SWAP]
sda
├─sda1 vfat   FAT32       728B-6CFB                               160M    69% /boot
├─sda2 ext4   1.0         cfb6037e-13c0-45f0-9962-127d985f14ce  130.4M    95% /
└─sda4 ext4   1.0         2cbfa280-60cd-43ab-aa7f-4fb68cc8380f    1.7G    97% /home
sdb
├─sdb1 ext4   1.0   data  3a1c94e3-a9c0-4579-b04e-1d9123efeadc
└─sdb2 vfat   FAT32 ESP   1906-770C

The Drive is sdb1 ofcourse
here is fsck: After I got a error

fsck from util-linux 2.38
e2fsck 1.46.5 (30-Dec-2021)
Pass 1: Checking inodes, blocks, and sizes
Inode 55705703 extent tree (at level 1) could be narrower.  Optimize<y>? yes
Pass 1E: Optimizing extent trees
Pass 2: Checking directory structure
Entry 'mmc-pack.json' in /multimc-instances/ME-instances/1.19 (38019100) has deleted/unused inode 38015482.  Clear<y>? yes
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
Block bitmap differences:  -(4030996--4030998) -4063256 +(4097572--4097573) +4163130 -4163626 -4163646 -(4164609--4164610) -222102111
Fix<y>? yes
Free blocks count wrong for group #123 (31827, counted=31830).
Fix<y>? yes
Free blocks count wrong for group #124 (32742, counted=32743).
Fix<y>? yes
Free blocks count wrong for group #125 (31695, counted=31693).
Fix<y>? yes
Free blocks count wrong for group #127 (3946, counted=3949).
Fix<y>? yes
Free blocks count wrong for group #4054 (325, counted=324).
Fix<y>? yes
Free blocks count wrong (126642452, counted=126642456).
Fix<y>? yes
Inode bitmap differences:  -(38015482--38015483)
Fix ('a' enables 'yes' to all) <y>? yes
Free inodes count wrong for group #4640 (3035, counted=3037).
Fix ('a' enables 'yes' to all) <y>? yes
Free inodes count wrong (60712788, counted=60712790).
Fix ('a' enables 'yes' to all) <y>? yes

data: ***** FILE SYSTEM WAS MODIFIED *****

      309418 inodes used (0.51%, out of 61022208)
        2692 non-contiguous files (0.9%)
         167 non-contiguous directories (0.1%)
             # of inodes with ind/dind/tind blocks: 0/0/0
             Extent depth histogram: 306534/711/1
   117416936 blocks used (48.11%, out of 244059392)
           0 bad blocks
          24 large files

      273439 regular files
       27893 directories
           0 character device files
           0 block device files
           0 fifos
           0 links
        8077 symbolic links (2164 fast symbolic links)
           0 sockets
------------
      309409 files

I replaced the cable with the one on the SSd and still there were error. I also did libata.force=5:3.0G
It didn't help

seth · 2022-06-19 11:42:47

You could check whether it's stress on the bus or elsewhere

dd if=/dev/zero of=/mnt/sdb1/path/to/wherever/foo.file count=2M # writes 1GB

If that doesn't cause any bus erros, try to do the same while stressing the CPU and/or GPU a bit, https://wiki.archlinux.org/title/Stress_testing

Are there any FW updates available for your MoBo?

Arch Linux

#1 2022-06-17 10:02:53

ata errors when under stress?

#2 2022-06-17 13:48:46

Re: ata errors when under stress?

#3 2022-06-18 03:23:23

Re: ata errors when under stress?

#4 2022-06-18 06:37:29

Re: ata errors when under stress?

#5 2022-06-19 10:25:47

Re: ata errors when under stress?

#6 2022-06-19 11:42:47

Re: ata errors when under stress?

Board footer