Odroid M1 (8GB) with debian 11 crashed just after few hours

[jazzLV]

Hi, everyone.
My brand new Odroid-M1 (8GB) with Debian 11 + openmediavault 6.3.2-2 crashed just few hours after installing (official method via petitboot netboot).
I went to sleep, left M1 powered ON and when woke up nothing worked - ssh failed to connect, openmediavault failed to login, samba shares were not working, etc. Hooked up monitor to M1 and was greeted by errors shown in picture:



Quick googling says its possible (known) bug (in Ubuntu at least) with some ssd's crashing when returning to sleep power states or something like that.
No idea if that is so..after re-powering device everything went back up and seems to work properly. Afterwards, I did uninstalled openmedivault-flashmemory plugin which was installed by default.
Somehow I suspect it might be responsible. But it just a speculation... I am using official power adapter 12V, 2A. So should not be problem with that.

If M1 will crash again this night or soon after I will remove openmediavault completely and reinstall debian 11 "from scratch". If it will crash even then, then will try Ubuntu.
My intent is to run M1 24/7, of course. Crashing is not acceptable...

m.2 ssd I am using is brand new Corsair Force MP510 NVMe PCIe Gen3 x4 M.2 1920GB SSD (CSSD-F1920GBMP510), firmware - ECFM15.1.
Also there is some old (but not really frequently used) Kingston microSD 64GB card inside M1 microsd slot (using exfat system).

p.s. I can see in openmediavault that everything seems normal in system (temps & load average) and everything seemed normal before crash too.

[odroid]

Did you try running 'fsck' to check your EXT4 file system integerity?

BTW, which kernel version do you run?
Did you install this @tobetter's Debian 11 OS?
viewtopic.php?f=52&t=40761

[jazzLV]

Yes, I am using tobetter's Debian 11 using netboot from petitboot.

Code: Select all

me@odroid:~$ cat /proc/version
Linux version 5.18.0-odroid-arm64 (tobetter@gmail.com) (gcc-10 (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.2) #1 SMP PREEMPT Debian 5.18.0-202205171802~bullseye (2022-05-17)

I did not run 'fsck'. My knowledge about debugging issues like this is pretty limited. Can you point me in right direction what exactly I should type for checking?
Do I need to run it before debian booted (in petitboot's shell)? And how long it might take to fully check 1.92TB nvme ssd ?

Here is output from my Debian 11:

Code: Select all

me@odroid:~$ df -h
Filesystem      Size  Used Avail Use% Mounted on
udev            3.7G     0  3.7G   0% /dev
tmpfs           745M  3.1M  742M   1% /run
/dev/nvme0n1p2  1.8T   41G  1.6T   3% /
tmpfs           3.7G     0  3.7G   0% /dev/shm
tmpfs           5.0M  4.0K  5.0M   1% /run/lock
tmpfs           3.7G  2.9M  3.7G   1% /tmp
/dev/mmcblk1p1   60G  2.4M   60G   1% /srv/dev-disk-by-uuid-5ECB-2E29
/dev/nvme0n1p1  472M   32M  417M   7% /boot
tmpfs           745M     0  745M   0% /run/user/1000

/dev/mmcblk1p1 is kingston's microSD card.

Code: Select all

me@odroid:~$ df -Th
Filesystem     Type      Size  Used Avail Use% Mounted on
udev           devtmpfs  3.7G     0  3.7G   0% /dev
tmpfs          tmpfs     745M  3.1M  742M   1% /run
/dev/nvme0n1p2 ext4      1.8T   41G  1.6T   3% /
tmpfs          tmpfs     3.7G     0  3.7G   0% /dev/shm
tmpfs          tmpfs     5.0M  4.0K  5.0M   1% /run/lock
tmpfs          tmpfs     3.7G  2.9M  3.7G   1% /tmp
/dev/mmcblk1p1 exfat      60G  2.4M   60G   1% /srv/dev-disk-by-uuid-5ECB-2E29
/dev/nvme0n1p1 ext2      472M   32M  417M   7% /boot
tmpfs          tmpfs     745M     0  745M   0% /run/user/1000

Btw, system did not crash after that 1 time (so far, so good).
Uptime right now is ~19 hours.

[odroid]

Googling with "fsck" to check the file system integrity once you meet a similar kernel message.

@tobetter,
Can we recommend to use Kernel 5.15 LTS or 6.1 LTS instead of 5.18 for a use case of personal NAS/OMV?
I think any kernel version might be fine if there is no stability issue though.

[jazzLV]

Managed to crash system after ~8 days uptime just by playing Bluray video (m2ts file) using PotPlayer from Windows 10 PC via local samba share from odroid (Gbit LAN).
It was playing nicely, but when I started moving playtime (basically "browsing" video to check its content) to various places in video and it eventually crashed odroid.
When attached monitor and keyboard to odroid there was blank screen.

Plugging OFF power source resulted in showing petitboot, then some disk checks (was too fast, did not get that on picture) and normal boot to debian.
I guess I will be more careful next time.

Some details about file that did that:

Code: Select all

Complete name  - 00001.m2ts
Format                                   : BDAV
Format/Info                              : Blu-ray Video
File size                                : 17.1 GiB
Overall bit rate                         : 34.2 Mb/s
Maximum Overall bit rate                 : 35.5 Mb/s

Still no idea how to check disk(s) integrity and at what stage should I do them. Was hoping for more specific replies for use case in odroid M1, not just generic "search google".
Thanks anyway that this forum exists.

[jazzLV]

Also discovered that my Odroid M1 crashes at some point when copying ~400GB total data (with files from ~1gb size to 10gb size) from odroid to PC via samba via gbit lan. At 23% it did crash. m2.ssd temp was rising quite quickly from 50C to 81C. Also I'm not sure could openmediavault would be also responsible for this crash somehow. Maybe without it (with just debian 11) it would perform more stable.

p.s. Additional issue - that after crash when there only thing to do is re-plug power source, network did not come back automatically.
Had to issue commands "dhclient -c -r" and then "dhclient -v" to get local IP from router.