[ @hdd @ssd @nvme @smart @checkdisk ]
Это “File System Consistency Check“, утилита для условного «лечения» файловой системы на HDD.
- Теория: Common:HowTo:Проверить:S.M.A.R.T.
- Диагностика: SmartMonTools
Для однозначности: fsck может условно «вылечить» постаревший HDD, но всегда разумнее просто заменить устаревший HDD на новый.
How to Use fsck Command to Check and Repair Filesystem
https://phoenixnap.com/kb/fsck-command-linux
Для устранения проблем fs не должна быть смонтирована на запись, то есть проверить на штатно загруженном linux корневые разделы на рабочем диске не выйдет. Надо загрузится не в эту ОС (с CD/DVD, флешки) и уже с неё смонтировать и проверить.
Есть программа badblocks - она проверяет именно диск на наличие сбойных блоков.
How to use fsck command to check and repair filesystem in Linux
Prerequisites
Access to a terminal or command line
A user with root permissions to run the tool
When to Use fsck in Linux
Note: Another useful tool for addressing a boot failure is GRUB rescue. Learn more in our post How to Use Grub Rescue to Fix Linux Boot Failure.
Basic fsck Syntax
The basic syntax for the fsck utility follows this pattern:
fsck <options> <filesystem>
In the above example, filesystem can be a device, a partition, a mount point, etc. You can also use filesystem-specific options at the end of the command.
How to Check and Repair Filesystem
There are a few steps to do before you check and repair your file system. You need to locate a device and unmount.
View Mounted Disks and Partitions
To view all mounted devices on your system and check disk location, use one of the available tools in Linux.
One method to locate the disk you want to scan is to list the filesystem disks with the df command:
df -h
Terminal output when running a df -h command
The tool prints the data usage on your system and filesystems. Take note of the disk you want to check with the fsck command.
To view partitions for your first disk, for example, use the following command:
sudo parted /dev/sda 'print'
sda is how Linux refers to your first SCSI disk. If you have two, the second would be sdb, and so on.
In our example, we got one result since there was only one partition on this virtual machine. You will get more results if you have more partitions.
The terminal output when listing Linux partitions.
The disk name here is /dev/sda and then the number of partitions is shown in the Number column. In our case, it is one: sda1.
Unmount the Disk
A warning when you try to unmount a mounted disk or partition.
Before you can run a disk check with fsck
you need to unmount a disk or partition. If you try to run fsck on a mounted disk or partition, you will get a warning:
Make sure to run the unmount command:
sudo umount /dev/sdb
Replace /dev/sdb with the device you want to unmount.
Note that you cannot unmount root filesystems. Hence, now fsck can’t be used on a running machine. More on that towards the end of the guide.
Run fsck to Check for Errors
Now that you unmounted the disk, you can run fsck. To check the second disk, enter:
sudo fsck /dev/sdb
The output after running a fsck command to check the second disk.
The above example shows the output for a clean disk. If there are multiple issues on your disk, a prompt appears for each one where you have to confirm the action.
The exit code the fsck utility returns is the sum of these states:
Possible exit codes for the fsck command.
Mount the Disk
When you finish checking and repairing a device, mount the disk so you can use it again.
In our case, we will remount the sdb disk:
mount /dev/sdb
Do a Dry Run with fsck
Before you perform a live check, you can do a test run with fsck. Pass the -N option to the fsck command to perform a test:
sudo fsck -N /dev/sdb
The output prints what would happen but does not perform any actions.
Fix Detected Errors Automatically with fsck
To try to fix potential problems without getting any prompts, pass the -y option to fsck.
sudo fsck -y /dev/sdb
This way, you say “yes, try to fix all detected errors” without being prompted every time.
If no errors are found, the output looks the same as without the -y option.
Skip Repair but Print fsck Errors in the Output
Use the -n option if you want to check potential error on a file system without repairing them.
We have a second drive sdb with some journaling errors. The -n flag prints the error without fixing it:
sudo fsck -n /dev/sdb
Use -n fsck option to print errors without fixing them.
Force fsck to Do a Filesystem Check
When you perform a fsck on a clean device, the tool skips the filesystem check. If you want to force the filesystem check, use the -f option.
sudo fsck -f /dev/sdc
Force the fsck tool to do a filesystem check
The scan will perform all five checks to search for corruptions even when it thinks there are no issues.
Bad magic number in super-block
Сначала определите, где находятся резервные копии суперблока. Это можно сделать с помощью команды mke2fs в режиме отображения резервных суперблоков (не создавая файловую систему):
sudo dumpe2fs /dev/sdc1 | grep -i superblock
Эта команда отобразит список резервных суперблоков, например:
Primary superblock at 0, Group descriptors at 1-699
Backup superblock at 32768, Group descriptors at 32769-33467
Backup superblock at 98304, Group descriptors at 98305-99003
Backup superblock at 163840, Group descriptors at 163841-164539
Backup superblock at 229376, Group descriptors at 229377-230075
Backup superblock at 294912, Group descriptors at 294913-295611
Backup superblock at 819200, Group descriptors at 819201-819899
Backup superblock at 884736, Group descriptors at 884737-885435
Backup superblock at 1605632, Group descriptors at 1605633-1606331
Backup superblock at 2654208, Group descriptors at 2654209-2654907
Backup superblock at 4096000, Group descriptors at 4096001-4096699
Backup superblock at 7962624, Group descriptors at 7962625-7963323
Backup superblock at 11239424, Group descriptors at 11239425-11240123
Backup superblock at 20480000, Group descriptors at 20480001-20480699
Backup superblock at 23887872, Group descriptors at 23887873-23888571
Backup superblock at 71663616, Group descriptors at 71663617-71664315
Backup superblock at 78675968, Group descriptors at 78675969-78676667
Backup superblock at 102400000, Group descriptors at 102400001-102400699
Backup superblock at 214990848, Group descriptors at 214990849-214991547
Backup superblock at 512000000, Group descriptors at 512000001-512000699
Backup superblock at 550731776, Group descriptors at 550731777-550732475
Backup superblock at 644972544, Group descriptors at 644972545-644973243
Запустите e2fsck с резервным суперблоком
Попробуйте восстановить файловую систему, указав первый резервный суперблок (например, 32768):
sudo e2fsck -b 32768 /dev/sdc1
Вероятне всего, по данному адресу будет найдено какое-то отклонение от нормы, которое может быть исправлено. Дальше можно нажимать клавишу y для каждого раза. Проще нажать клавишу a, чтобы автоматически соглашаться с каждой следующей проверкой.
Терпеливо подождать, пока всё отработает, на современных дисках это длится долго.
Run fsck on All Filesystems at Once
If you want to perform a check on all filesystems with fsck in one go, pass the -A flag. This option will go through the etc/fstab file in one run.
Since root filesystems can’t be unmounted on a running machine, add the -R option to skip them:
fsck -AR
To avoid the prompts, add the -y option we talked about.
Skip fsck on a Specific Filesystem
If you want fsck to skip checking a filesystem, you need to add -t and “no” before a filesystem.
For example, to skip ext3 filesystem, run this command:
sudo fsck -AR -t noext3 -y
We added -y to skip the prompts.
Skip Fsck on Mounted Filesystems
To make sure you do not try to run fsck on a mounted filesystem, add the -M option. This flag tells the fsck tool to skip any mounted filesystems.
To show you the difference, we will run fsck on sdb while it is mounted, and then when we unmount it.
sudo fsck -M /dev/sdb
The output of fsck tool to skip any mounted filesystems.
While sdb is mounted, the tool exits without running a check. Then, we unmount sdb and run the same command again. This time, fsck checks the disk and reports it as clean, or with errors.
Note: To remove the first title line of the fsck tool “fsck from util-linux 2.31.1” use the -T option.
Run fsck on Linux Root Partition
As we already mentioned, fsck cannot check root partitions on a running machine since they are mounted and in use. However, even Linux root partitions can be checked if you boot into recovery mode and run the fsck check:
- To do so, power on or reboot your machine through the GUI or by using the terminal:
sudo reboot
- Press and hold the shift key during boot-up. The GNU GRUB menu appears.
- Select Advanced options for Ubuntu.
Linux recovery mode screen.
- Then, select the entry with (recovery mode) at the end. Let the system load into the Recovery Menu.
- Select fsck from the menu.
Linux recovery menu select fsck tool.
- Confirm by selecting <Yes> at the prompt.
Recovery mode confirmation message when fsck is selected.
- Once finished, select resume at the recovery menu to boot up the machine.
Resume option selected when the tool finishes checking a root partition.
What if fsck is Interrupted?
You should not interrupt the fsck tool while it is in progress. However, if the process is interrupted, fsck will finish the ongoing check and then stop.
In case the utility found an error while the check was in process, it will not try to fix anything if interrupted. You can rerun the check next time and let it finish.
fsck Linux Command Options Summary
To wrap up, below is the list of the options you can use with the fsck Linux utility.
Option Description
-a Try to repair filesystem errors automatically. There will be no prompts, so use it with caution.
-A Check all filesystems listed in /etc/fstab.
-C Show progress for ext2 and ext3 filesystems.
-f Force fsck to check a filesystem. The tool checks even when the filesystem appears to be clean.
-l Lock the device to prevent other programs from using the partition during the scan and repair.
-M Do not check mounted filesystems. The tool returns an exit code 0 when a filesystem is mounted.
-N Do a dry run. The output prints what the fsck would do without executing any actions. The warning or error messages are printed as well.
-P Use to run a scan on multiple filesystems in parallel. It can cause issues, depending on your setup. Use with caution.
-R Tell the fsck tool not to check the root filesystems when you use the -A option.
-r Print device statistics.
-t Specify which filesystems type(s) to check with fsck. Consult the man page for detailed information.
-T Hide the title when the tool starts.
-y Try to repair filesystem errors automatically during the check.
-V Verbose output.
Note: Learn about the error code SIGSEGV (signal segmentation violation) and how to troubleshoot it.