LVM logical volume for /home partition corrupted after lvresize to increase swap space Fedora 25


I added 6GB of memory on an older Dell PowerEdge 1850 that we use just for backups to an attached hardware RAID from Promise Technology called a VessRAID connected by iSCSI. After a few weeks I noticed the swap space was only 2 GB and decided to resize the home partition and add 6GB to swap.

I followed the steps to Extend Swap on an LVM2 Logical Volume. On reboot I got to emergency mode and the dreaded and "Either the superblock or the partition table is likely to be corrupt!". I also see failed to find cpu0 device node error. The server is on the 4.9.3 kernel. Is my best bet still to try mke2fs with or without the -S option as suggested here? I also tried the resize2fs command but I get "Device or resource busy while trying to open /dev/sdb"

Not sure if it's related but here are some other errors I was seeing in the logs before reboot:

Jan 19 20:06:00 curie systemd-logind: Removed session 156.
Jan 19 20:14:56 curie systemd: Failed to set up mount unit: Device or resource busy
Jan 19 21:10:18 curie systemd: Cannot add dependency job, ignoring: Unit home.mount is masked.
Jan 19 21:11:14 curie systemd: Cannot add dependency job, ignoring: Unit -.mount is masked.

/etc/fstab has this:

/dev/mapper/vg_curie-lv_root /                       ext4    defaults        1 1
UUID=e3507fa1-f9ee-4951-8ecc-48f16183dee9 /boot                   ext4    defaults        1 2
/dev/mapper/vg_curie-lv_home /home                   ext4    defaults        1 2
/dev/mapper/vg_curie-lv_swap swap                    swap    defaults        0 0

Best Answer

Well I got lucky. These commands seem to have worked!

mkfs.ext4 -S /dev/mapper/vg_curie-lv_home 
e2fsck -f -y /dev/mapper/vg_curie-lv_home 

The fsck took a couple of hours but it came back with all errors fixed and a reboot worked!