IJ26788: A "SILENT" FAILURE TO INIT DISK FENCING LEADS TO RG ERROR LATER

APAR status

• Closed as program error.

Error description

• A temporary issue which leads to a
  failure of clpr_verifycap_vg and/or
  clpr_reg_res_vg (initialization of
  disk fencing) during a node_up event,
  e.g.:
  :node_up[node_up_scsipr_init:68]
   clpr_verifycap_vg datavg011
  cl_scsipr_verifycap: "openx(/dev/hdisk5,
    O_RDONLY, S_IRUSR, SC_NO_RESERVE)"
    failed with errno 13
  
  should be handled other than just to
  print a notification in hacmp.out:
  :node_up[node_up_scsipr_init:166]
   : VG [shared VG] is not SCSIPR Type 7 capable.
  :node_up[node_up_scsipr_init:167]
   : [shared VG] will not be varied on and
     corresponding RG will go to error state.
  :node_up[node_up_scsipr_init:168]
   : There are one or more disks in the VG [shared VG]
     which are not SCSIPR capable.
  :node_up[node_up_scsipr_init:169]
   : Select the disks which are SCSIPR Type 7 capable
     or disable Disk Fencing
  :node_up[node_up_scsipr_init:170]
   : in order bring this Volume Group online.
  
  If no RG is acquired as part of the PowerHA
  startup, the failure does not lead to an
  event error, but (maybe much) later to
  an error acquiring the RG.
  

Local fix

Problem summary

• A resource group fails to acquire on a node if its volume group
  disks do not have SCSI PR key registration. This problem is seen
   when PR key registration fails on a remote node during cluster
  start operation and this will not be notified to user except a
  log message in hacmp.out
  

Problem conclusion

• A new check has been added in cluster auto verification process
  that notifies user with the problem and a corrective action
  

Temporary fix

• NA
  

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

  IJ29548 IJ31618 IJ31619

Fix information

• Fixed component name

  POWERHA SYSMIR

• Fixed component ID

  5765H3900

Applicable component levels