[zfcp] SCSI paths to IBM 2145 LUNs failing repeatedly

Issue

• SCSI paths to IBM 2145 LUNs connected through zfcp are failing repeatedly:

  multipathd: mpatha: sdad - tur checker reports path is down
  multipathd: mpatha: sdaj - tur checker reports path is down
  multipathd: sdaj: couldn't get asymmetric access state
  multipathd: mpathb: sdck - tur checker reports path is down
  multipathd: mpathb: sdcl - tur checker reports path is down
  [...]
  
  mpatha (3600517181110111c10101010101016121) dm-3 IBM,2145
  size=500G features='1 queue_if_no_path' hwhandler='0' wp=rw
  `-+- policy='round-robin 0' prio=30 status=active
    |- 0:0:1:5 sdl  8:176  active ready running
    |- 0:0:2:5 sdf  8:80   active ready running
    |- 1:0:2:5 sdr  65:16  active ready running
    |- 1:0:1:5 sdx  65:112 active ready running
    |- 2:0:0:5 sdad 65:208 failed faulty running  <---
    |- 2:0:1:5 sdaj 66:48  failed faulty running  <---
    |- 3:0:2:5 sdap 66:144 active ready running
    |- 3:0:1:5 sdav 66:240 active ready running
    |- 4:0:3:5 sdbb 67:80  active ready running
    |- 4:0:2:5 sdbh 67:176 active ready running
    |- 5:0:1:5 sdbn 68:16  active ready running
    `- 5:0:3:5 sdbt 68:112 active ready running
  

• These repeated path failures also causes IO hang, and performance issues.