Filesystem and fstab errors can prevent boot or cause data loss. Always run findmnt --verify and keep a rescue environment available.
What you will achieve
Recover from read-only root remounts after I/O errors or full disks — remount rw safely and fix underlying cause.
1) Confirm read-only state
mount | grep ' / '
touch /tmp/test 2>&12) Check kernel log for I/O errors
sudo dmesg | tail -30
journalctl -k -b | grep -i error3) Remount read-write (if disk is healthy)
sudo mount -o remount,rw /
If this fails, boot single-user or live USB and run fsck.
4) Filesystem check offline
sudo systemctl emergency
sudo umount /
sudo fsck -f /dev/sda1Replace device with your root partition from findmnt /.
5) Disk full cause
df -h
sudo du -xh --max-depth=1 /var | sort -hVerify
mount | grep ' / '
sudo touch /etc/ro-test && sudo rm /etc/ro-test5) ext4 errors in dmesg
sudo dmesg | grep -i ext4ABORT or I/O error messages mean hardware or cable — replace disk after backup.
6) Read-only bind mounts
findmnt -o TARGET,OPTIONS /Look for ro flag — intentional ro mount differs from emergency remount.
7) Live USB fsck when root cannot remount rw
sudo fsck.ext4 -f /dev/sda2Prevent recurrence
Monitor disk health, keep 15% free space, use UPS on servers — most ro remounts are the kernel protecting data on sick disks.
8) OverlayFS live USB false positive
Live sessions mount squashfs ro by design — do not fsck the ISO stick thinking it is a broken HDD.
Prerequisites
Root access or single-user mode. Know root device (findmnt /). Backup if disk failing. Maintenance window — remount and fsck may require downtime.
Emergency read-only root cause log
journalctl -k -b | grep -i 'read-only\|I/O error\|EXT4'smartctl long test
sudo smartctl -t long /dev/sda
sudo smartctl -a /dev/sdaSchedule long SMART test when intermittent ro remounts suggest failing sectors.
Network filesystem ro
NFS server export disappeared mid-write — client mounts ro and apps crash. mount -o remount,rw fails until server returns — fix server not local fsck. Same for stale SSHFS when remote host down.
Container overlay ro
Docker overlay2 ro often means underlying host disk full — fix host df not container restart loop. kubectl evicting pods when node filesystem read-only — drain node and fix disk on host.