EXT4-fs issue after moving VM to a new host

[simplerezo]

Hi !

After moving a VM from an host to another one, a working VM Ubuntu is failing randomly (sometime just 5 minutes after booting, even with few activity) :

sept. 11 10:33:34 XXX kernel: EXT4-fs error (device sda3): ext4_validate_inode_bitmap:105: comm less: Corrupt inode bitmap - block_group = 13344, inode_bitmap = 437256208
sept. 11 10:33:34 XXX kernel: Aborting journal on device sda3-8.
sept. 11 10:33:34 XXX kernel: EXT4-fs (sda3): Remounting filesystem read-only

This requires a reboot and a fsck...

On the host (using proxmox8 and ZFS) there is no error reported.

Since the host is using NVMe disks, we tried to add "nvme_core.default_ps_max_latency_us=0" to kernel cmdline (on host and in VM), but this does not make any difference... (I know this should not be required anymore)

Config infos:
- host : proxmox 8 / ZFS / nvme disk
- old host : proxmox 8 / ZFS / SSD disks
- vm : ubuntu 23.04 / VirtIO SCSI single

 

[simplerezo]

Error is not always exactly the same:

[ 7819.506602] EXT4-fs error (device sda3): ext4_validate_block_bitmap:422: comm bash: bg 610: bad block bitmap checksum
[ 7819.506640] Aborting journal on device sda3-8.
[ 7819.507909] EXT4-fs error (device sda3): ext4_journal_check_start:83: comm rs:main Q:Reg: Detected aborted journal
[ 7819.507916] EXT4-fs error (device sda3): ext4_journal_check_start:83: comm systemd-journal: Detected aborted journal
[ 7819.512136] EXT4-fs (sda3): Remounting filesystem read-only
[ 7819.512155] EXT4-fs error (device sda3) in ext4_do_writepages:2890: IO failure