Fresh install system is not usable

[utamav]

On a fresh install of OpenSUSE 16.0, the system goes in a state where most commands don’t work. It’s installed as a server.

> fsck -f /dev/sde
-bash: /usr/bin/cnf: Input/output error

> sudo fsck -f /dev/sde
sudo: unable to open /var/lib/sudo/ts/1000: Read-only file system
[sudo] password for user: 
Segmentation fault

> journalctl -xb
No journal files were found.
Failed to execute 'less', will try 'more' next: Input/output error

> df -kh
-bash: /usr/bin/df: Input/output error

A reboot usually resolves the issue but only for a couple of hours or minutes.

[shundhammer]

This might be a legitimate hardware problem with your disk or controller, or an SELinux problem with the SELinux filesystem labelling where the “security contexts” are set up for each directory (and file? Not sure).

See also

• https://en.opensuse.org/Portal:SELinux
• https://en.opensuse.org/Portal:SELinux/Tools
• https://doc.opensuse.org/documentation/leap/security/html/book-security/cha-selinux.html
• https://docs.redhat.com/en/documentation/red_hat_enterprise_linux/7/html/selinux_users_and_administrators_guide/sect-security-enhanced_linux-working_with_selinux-selinux_contexts_labeling_files

[utamav]

I thought it was SELinux related and that is why I did a fresh install without SELinux enabled this time but seems to be the same result. I’ll try a different boot drive to see if the problem is hardware related.

[utamav]

I don’t believe it’s hardware. I tried another external SSD drive with a fresh install. Similar results in less than 24 hours.

[deano_ferrari]

I recommend booting a live Linux distro to test the RAM, CPU, and storage controller. You can also try the built-in UEFI diagnostic tests to help identify any hardware issues.

You could run a stress test via the live distro eg

sudo dd if=/dev/zero of=/dev/sde bs=1M count=1024 oflag=direct status=progress
sudo dd if=/dev/sde of=/dev/null bs=1M iflag=direct status=progress

This is a destructive test so make sure you write to the correct device under test.

A safe (nondestructive) SMART self-test, assuming that the live dsitro has this utility included…

sudo smartctl -a /dev/sde       # View full SMART status
sudo smartctl -t short /dev/sde # Run a short self-test
sudo smartctl -t long /dev/sde  # Run a full self-test (may take hours)

Running the live distro on the system under test may also yield dmesg output that can help determine if there is an underlying issue with the host system.