Comment 52 for bug 1567602

------- Comment From <email address hidden> 2016-06-28 08:15 EDT-------
(In reply to comment #60)
> Hi Thorsten, so maybe (if the main issue is the inquiry) it is possible to
> get away with just picking the change that appears to remove that in favour
> of present device information. Could you try the kernel at
> http://people.canonical.com/~smb/lp1567602/ in your environment and let me
> know whether that helps anything?

Please excuse the late response. That update slipped through :-(

Nevertheless, behaviour is slightly better, but does not fix the problem completely nor gets rid of the described symptoms:
I installed the above mentioned kernel manually, loaded scsi_dh_alua kernel module and set two zfcp devices online (which will auto-enable four LUNs each) via chccwdev -e 1904,1944. Result is:
Sometimes the LUN detection runs through until all four LUNs are detected, sometimes it terminates LUN detection at an earlier time. In most cases the correlation is as follows: for the first zfcp device, 0...2 LUNs are detected, for the second zfcp device of the 'chccwdev -e' sequence 3...4 LUNs are detected.

So: Yes, it helps something. But it ist not completely fixed (nor worked around).