"blk_update_request: I/O error, dev nbdN" follows "Connection timed out" on a third-party NBD device

Issue

• I/O errors for an nbd device (e.g., nbd1) are recorded in the kernel log (/var/log/messages or journalctl -k). These errors may be preceded by messages suggesting an NBD connection timeout or failure.
• What is the cause of this error and what action is required?
• Example Log Sequence:

kernel: block nbd1: Connection timed out, retrying (1/1 alive)
kernel: block nbd1: Receive control failed (result -32)
kernel: block nbd1: Dead connection, failed to find a fallback
kernel: block nbd1: shutting down sockets
kernel: blk_update_request: I/O error, dev nbd1, sector 911608 op 0x1:(WRITE) flags 0x800 phys_seg 6 prio class 0
kernel: md/raid1:md0: Disk failure on nbd1, disabling device.#012md/raid1:md0: Operation continuing on 1 devices.
nbd-client[511357]: Kernel call returned.
nbd-client[511357]: Begin Closing
nbd-client[511357]: Closing Complete