How to run fsck in case of SSB File System Corruption (4262752)

• The order of the steps are strict! Please keep in mind, taking them in a different order could cause message loss and even an unbootable system!
• You have to have direct console access or access to the IPMI, because during the process the network connection will be down.
• This procedure applies to the file system of the core firmware, but not the system's root file system.
• In case of a cluster, the slave node must be stopped before the procedure

1. Login to the boot shell via IPMI or virtual console.

2. Get the device name of the core firmware, that will be used in step 5.

mount |awk '/\/mnt\/firmware\/mnt\/firmware/ { print $1 }'

3. Stop the core firmware:

systemctl stop boot-xcb

4. Only appliances with DRBD (output of step 2 is /dev/drbd0), set the DRBD device to 'primary':

drbdsetup /dev/drbd0 primary

5. Issue fsck, fsck.ext4 -f -v DEVICE_FROM_STEP_2

Apliances with DRBD

fsck.ext4 -f -v /dev/drbd0

Virtual appliances installed SSB 5 LTS or later

fsck.ext4 -f -v /dev/sda4

6. Exit from the boot shell and reboot the node:

Control -> Reboot