View source for Either the superblock or the partition table is likely to be corrupt fix

If your server show the "the superblock or the partition table is likely to be corrupt " error during the boot/fsck this may cause the physical problem but not for sure.

Such issue may occur after  incorrect partition/LVM resizing /reducing . 
It may happens even few days after shrinking. 

The partition mounted and all  data was present. However the next day/boot when the file systems were being checked at boot time, shrunk file system failed with this error:

 The filesystem size (according to the superblock) is 10585760 blocks
 The physical size of the device is 10577568 blocks
 Either the superblock or the partition table is likely to be corrupt!

As you can see the physical device size is less than filesystem size. So you need to set filesystem size less or equal to physical:

 resize2fs /dev/<dev> 10577568
Replace <dev> with proper device, for example [we are working with LVM] :
 resize2fs /dev/vg_kvm_gr/LogVol001 10577568

Wait some time. Once finished check with fsck or e2fck:
 fsck -f /dev/vg_kvm_gr/LogVol001

The issue should be fixed and you should be able to mount device or simply reboot server. If the issue still exists  - probably you need to check physical state of your disks .
Anyway we suggest to complete the full backup using dd or any other utility

'''If you are using LVM you may note that the issue appeared inside the kvm virtual servers that located on the already fixed lvm volume.

Sometimes the guest OS do not able to even detect the disk.

First of all try to disable virtio and writeback caching.  Reboot VM, after that use fsck or chkdsk inside the container.

'''Also if you are using some software raid like mdadm check the raid status and run manual resync in case it neccessary.
Tip: take a look on the io load (use atop or iotop) and make proper sync after ..like sdb->sda or vice versa

[[Category:Admin area]]