nvme-tcp automatic reconnect fails intermittently during EMC PowerStore NDU operation

Issue

• nvme-tcp automatic reconnect fails intermittently during EMC PowerStore NDU operation:

kernel: nvme nvme45: Connect command failed, error wo/DNR bit: -16389
kernel: nvme nvme45: failed to connect queue: 9 ret=-5
....
kernel: nvme nvme49: Failed reconnect attempt 75
kernel: nvme nvme49: Reconnecting in 10 seconds...
kernel: nvme nvme45: queue_size 128 > ctrl sqsize 64, clamping down
kernel: nvme nvme47: queue_size 128 > ctrl sqsize 64, clamping down
kernel: nvme nvme46: queue_size 128 > ctrl sqsize 64, clamping down
kernel: nvme nvme50: queue_size 128 > ctrl sqsize 64, clamping down
kernel: nvme nvme45: creating 32 I/O queues.
kernel: nvme nvme53: queue_size 128 > ctrl sqsize 64, clamping down
kernel: nvme nvme49: queue_size 128 > ctrl sqsize 64, clamping down
kernel: nvme nvme46: creating 32 I/O queues.
kernel: nvme nvme50: creating 32 I/O queues.
kernel: nvme nvme47: creating 32 I/O queues.
kernel: nvme nvme53: creating 32 I/O queues.
kernel: nvme nvme49: creating 32 I/O queues.
kernel: nvme nvme41: queue_size 128 > ctrl sqsize 64, clamping down
kernel: nvme nvme41: creating 32 I/O queues.
kernel: nvme nvme45: Connect command failed, error wo/DNR bit: -16389
kernel: nvme nvme45: failed to connect queue: 9 ret=-5
kernel: nvme nvme45: Failed reconnect attempt 76
kernel: nvme nvme45: Reconnecting in 10 seconds...
kernel: nvme nvme47: Connect command failed, error wo/DNR bit: -16389
kernel: nvme nvme47: failed to connect queue: 9 ret=-5
kernel: nvme nvme46: Connect command failed, error wo/DNR bit: -16389

• Occasionally, the automatic rediscovery on the nvme-tcp devices will fail. To correct, the end user will have to manually disconnect the controller and reconnect again.