FC 1242SR can't recover IO when plug fibre on diskarrays in 3 seconds

[JamesChao]

Hi:

Question descriped as bellow:

The FC Fabric includes a HP Server, a Brocade switch and a disk array. The server use a HP HBA FC 1242SR.

When the server is running IO to the disk array, I plug and insert the fibre on the disk array in 3s at most. I expect the IO

recovering smoothly. But it's stopped every time.

According the messages on disk array and the trace, the disk array recover the link to switch immediately, but the HBA sends

PLOGI to disk array almost 6s later. And the server prints messages like this:
"blocked FC remote port time out: removing target and saving binding".

I attempted to modify the HBA "Port Down Retry Count" parameter from 8 to 16 or 32, and the PLOGI delay time followed to

increase.

I guess it's a bug in XenServer 5.5, but I don't have enough evidence. Maybe you can help me to analyze this question.

Thank you!

PS:

The accessories is the server messages, HBA information and trace.

I used HBA port named 50-01-43-80-06-34-88-44.