dmp_fast_recovery is set to off
# vxdmpadm gettune dmp_fast_recovery
Tunable Current Value Default Value
------------------------------ ------------- -------------
dmp_fast_recovery off on
System messages shows reservation conflict messages
kernel: VxVM vxdmp V-5-0-0 SCSI error opcode=0x5f returned rq_status=0x12 cdb_status=0x1 key=0x6 asc=0x2a ascq=0x3 on path 8/0x90
kernel: sd 2:0:0:1: reservation conflict
kernel: sd 2:0:0:1: Unhandled error code
kernel: sd 2:0:0:1: SCSI error: return code = 0x00110018
kernel: Result: hostbyte=invalid driverbyte=DRIVER_OK,SUGGEST_OK
dmpevents.log shows I/O error for other paths
I/O error occurred on Path sdg belonging to Dmpnode emc0_03f8
I/O analysis done as DMP_PATH_OKAY on Path sdg belonging to Dmpnode emc0_03f8
Marked as ioerr Path sdg belonging to Dmpnode emc0_03f8
I/O error occurred on Path sdg belonging to Dmpnode emc0_03f8
Unmarked as ioerr Path sdg belonging to Dmpnode emc0_03f8
vxfenadm shows only 1 key is registered confirming missing keys for remaining paths.
# vxfenadm -s /dev/vx/dmp/emc0_03f8
Reading SCSI Registration Keys...
Device Name: /dev/vx/dmp/emc0_03f8
Total Number Of Keys: 1
key[0]:
[Numeric Format]: 66,86,67,83,0,0,0,0
[Character Format]: BVCS
[Node Format]: Cluster ID: unknown Node ID: 1 Node Name: testsys1
The issue is identified as a defect.
When dmp_fast_recovery is off, the reservation flag is not set properly, due to which the path registration for the remainder of the paths fails with the transport error and PGR keys fail to get registered. This results in subsequent reservation conflict messages.
But when dmp_fast_recovery is on ; the reservation flag is set correctly thereby preventing this issue.
This issue has been fixed in
1) 5.1 SP1 RP3.P1 onward
2) 6.0.3 onward
If the dmp_fast_recovery tunable is set to off, PGR (Persistent Group Reservation) key registration fails except for the first path i.e. only for the first path PGR key gets registered. Post that reservation conflict message is logged for the rest of the paths.
ETrack: 2928764