What this article covers
Explains how to identify onboard flash storage file system corruption, kernel read-only remount errors and boot block read errors on FortiGate appliances.
Before using the CLI
On-site workflow
Detect early flash failure: Zombie miglogd & 0WA I/O wait spike
Before flash storage permanently locks into read-only mode, failing eMMC/SSD flash exhibits distinct signatures: in diagnose sys top, miglogd transitions to State Z (Zombie) or D (uninterruptible sleep), CPU I/O wait (0WA) spikes dramatically (30% to 70%+), and Cached RAM inflates as write buffers stall.
diagnose sys top 2 50
# Check miglogd state (Z/D) and CPU 0WA percentage
diagnose hardware sysinfo memory | grep -i cached
diagnose sys disk status
Inspect crash, kernel remount & EXT4 block errors
Look for EXT4-fs errors (e.g. ext4_lookup), I/O errors (e.g. Buffer I/O error on device), and kernel messages remounting partitions read-only.
diagnose debug crashlog read
# Check for: EXT4-fs error or Remounting filesystem read-only
Validate the recovery command before use
The supplied source lists a filesystem scan or repair procedure using the following commands. Confirm the exact model, target disk and command effect with current documentation or TAC before execution.
execute formatlogdisk
execute disk scan
These operations may erase log storage, modify filesystems or require a reboot. Do not run them as routine diagnostics.
Use console and bootloader recovery when required
If the boot filesystem is corrupted, the supplied source directs the engineer to interrupt the BIOS boot sequence and re-flash FortiOS with TFTP from the bootloader menu.
Bootloader recovery is outage-causing and can remove the installed image or configuration. Use a verified image, console access and a rollback plan.
Escalate persistent read-only flash
When flash remains read-only after the supported recovery process, initiate an RMA request for hardware replacement.