How to determine what exactly is the problem with the disk? What tools to use?

To determine if the problem is specifically with the disk, you need to check several key points:

1. System logs and SMART status: Use dmesg, journalctl to look for disk errors (e.g., I/O errors). The command smartctl -a /dev/sdX will show the SMART status — presence of error sectors, reallocated sectors, etc.

2. Performance tests:

   • iostat — shows input/output statistics, delays, and disk load.
   • fio — for stress testing the disk.

3. Monitoring delays and errors:

   • vmstat and iostat will help identify high wait times for operations.
   • Check for errors in /var/log/syslog or /var/log/messages.

4. File system check:

   • fsck to verify integrity.

Example of SMART check:

smartctl -a /dev/sda

If read errors, reallocated sectors, or other warnings are visible — a disk problem is likely.