XFS filesystem shuts down due to SCSI command timeouts in a virtual environment

Issue

• An application running on a virtual machine becomes unresponsive.
• Upon checking the kernel logs (/var/log/messages or journalctl -k), a sequence of messages is found, starting with SCSI command timeouts and I/O errors, immediately followed by messages indicating that the XFS filesystem has shut down.

kernel:sd 0:0:1:0: [sdb] tag#632 timing out command, waited 180s
kernel:sd 0:0:1:0: [sdb] tag#632 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_OK cmd_age=180s
kernel:sd 0:0:1:0: [sdb] tag#632 CDB: Write(16) 8a 00 00 00 00 00 c0 00 cd 40 00 00 00 40 00 00
kernel:blk_update_request: I/O error, dev sdb, sector 3221278016 op 0x1:(WRITE) flags 0x1000 phys_seg 8 prio class 0
[...]
kernel:dm-0: writeback error on inode 2149684214, offset 11210752, sector 2152198048
kernel:XFS (dm-0): log I/O error -5
kernel:XFS (dm-0): xfs_do_force_shutdown(0x2) called from line 1250 of file fs/xfs/xfs_log.c. Return address = 000000000d2dcf25
kernel:XFS (dm-0): Log I/O Error Detected. Shutting down filesystem
kernel:XFS (dm-0): Please unmount the filesystem and rectify the problem(s)