Connectrix B-Series Hardware : Customer reported BB credits zero events (tim_txcrd_z) on switch port.

>>>>> Customer reported the tim_txcrd_z counter increasing on switch port.

>>>>> Confirmed the BB credits zero events on switch port:

>>>>> Port 31 stats showing BB credit zero events:
 
tim_txcrd_z         7002878     Time TX Credit Zero (2.5Us ticks)
tim_txcrd_z_vc  4- 7:  0           7002878     0           0 
 
tim64_txcrd_z 0           top_int : Time BB_credit zero
              7002878     bottom_int : Time BB_credit zero

>>>>> According to Brocade Command Reference Guide the definiton of tim_txcrd_z_vc counter:
 
The number of times that the port was unable to transmit frames because the transmit BB credit was zero for each of the port's 16 Virtual Channels (VC 0-15).
The purpose of this statistic is to detect congestion or a device affected by latency.
This parameter is sampled at intervals of 2.5 microseconds (microseconds), and the counter is incremented if the condition is true. Each sample represents 2.5 microseconds of time with zero Tx BB Credit.
An increment of this counter means that the frames could not be send to the attached device for 2.5 microseconds, indicating degraded performance (platform- and port-specific).

>>>>> Conclusion:
 
The switch and device negotiate the numbers of buffers which can be used for data transfer. Buffers mean in general memory per port to store data which have to be send.
 
There's a process that counts free buffers on a port per second. when there are zero free buffers, the counter is increased. In other words, if the counter increases, it means that all buffers are allocated during the fair amount of time. It doesn't mean there was no free buffer so it doesn't necessarily mean that there are not enough buffers. This is evidence that the switch transport side is waiting for the device which is connected to that port.
 
If the counter increase on a F-Port this means that the switch ports has send 8 data frames out and has not received any receiver-ready message back for more than 2.5us. The device is processing and have no time for IO requests currently. In example: not all HBA have the same amount of memory on board.
 
If you could increase the number of buffers on any F-Port on switch side, it would increase your problem on the connected device.
 
For ISLs you can configure Extended Fabric parameter which increase the number of BB credits on E_ports depending of the distance one switch is from the other.