[Symptom Description]
In a cluster service scenario, after a mapping view, such as the mapping view name, alias, and command device, is modified, the host cluster enters the recovery process, and upper-layer services are occasionally suspended for 30 seconds.
[Alarm Information]
NA.
[Cause Description]
When the mapping view information is modified, the LUN reservation information is cleared. After the LUN reservation information is cleared, the host cluster detects that the reservation is lost and starts the reservation restoration process. During this process, services are occasionally interrupted.
[Troubleshooting Roadmap]
Condition 1: The upper-layer host is a cluster service and uses SCSI reservation.
Condition 2: The mapping view is modified on the storage device, such as the name, alias, and command device.
Both the preceding conditions are met, and the following information is displayed on the LUN corresponding to the disk array log:
1) Search for log 1 "SYS_CFG: Receive client msg, opcode(55834583040), objType(245), name(modify mapview)."."
1) Search for the log 2 "TGT_LMP:Reservation is clear (ret -1, LUN 861).][log_tgt]".
Logs 1 and 2 exist at the time when the fault occurs, and the log interval is very short (less than 3 seconds).
[Emergency Measures]
NA: The cluster is automatically recovered.
[Post-Recovery Check]
N.A.
[Applicability]
Dorado V3R1C00,Dorado V3R1C01,Dorado V3R1C00SPC100,Dorado V3R1C01SPC100,Dorado V3R1C20,Dorado V3R1C21,Dorado V3R1C21SPC100,Dorado V3R1C30,Dorado V3R1C30SPC100,Dorado V3R2C00,Dorado V3R2C00SPC100
