Hi there,
I installed multipath in redhat 6.6 in Dell R730XD with 02 FC Emulex card.

My SAN storage is Hitachi. I mapped SAN successfuly but got error as bellow, and could not record data smothly.

This is the first time I work with multipath and don't know reason of issue.

Thank you for your support,

Oct 9 12:50:02 vaspro05 multipathd: mpathh: sdc - tur checker reports path is up

Oct 9 12:50:02 vaspro05 multipathd: 8:32: reinstated

Oct 9 12:50:02 vaspro05 multipathd: mpathh: remaining active paths: 2

Oct 9 12:50:02 vaspro05 multipathd: mpathh: switch to path group #1

Oct 9 12:50:16 vaspro05 kernel: scsi 2:0:1:0: rejecting I/O to offline device

Oct 9 12:50:16 vaspro05 kernel: scsi 2:0:1:0: rejecting I/O to offline device

Oct 9 12:50:16 vaspro05 kernel: scsi 2:0:1:0: rejecting I/O to offline device

Oct 9 12:50:19 vaspro05 multipathd: 8:32: mark as failed

Oct 9 12:50:19 vaspro05 multipathd: mpathh: remaining active paths: 1

Oct 9 12:50:19 vaspro05 kernel: sd 2:0:0:2: [sdc] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE

Oct 9 12:50:19 vaspro05 kernel: sd 2:0:0:2: [sdc] Sense Key : Aborted Command [current]

Oct 9 12:50:19 vaspro05 kernel: sd 2:0:0:2: [sdc] <<vendor>> ASC=0xc0 ASCQ=0x1ASC=0xc0 ASCQ=0x1

Oct 9 12:50:19 vaspro05 kernel: sd 2:0:0:2: [sdc] CDB: Write(10): 2a 00 00 01 98 60 00 00 20 00

Oct 9 12:50:19 vaspro05 kernel: end_request: I/O error, dev sdc, sector 104544

Oct 9 12:50:19 vaspro05 kernel: device-mapper: multipath: Failing path 8:32.

Oct 9 12:50:20 vaspro05 multipathd: mpathh: sdc - tur checker reports path is up

Oct 9 12:50:20 vaspro05 multipathd: 8:32: reinstated

Oct 9 12:50:20 vaspro05 multipathd: mpathh: remaining active paths: 2

Oct 9 12:50:20 vaspro05 multipathd: mpathh: switch to path group #1

I'd suggest you install lsscsi (yum install) then run that command to see what SCSI disks (including SAN attached) are seen by the OS.

I'd also suggest you install Hitachi's HORCM software for Linux and run "inqraid -CLI -fxng" command to see what disks it thinks are assigned to the host.

Often you have to rescan the fiber HBAs to make them take notice of drives you've mapped in. The two tools mentioned will let you know for sure that the disks you think are mapped are in fact seen and accessible by the host.

You didn't mention fiber switches for your SAN. Are you using any? Did you update zoning there to pass the disks through? Here we have map the drives both in our SAN switches and in our Hitachi array.

For RHEL6 make sure you've updated to the most recent device-mapper-multipath and device-mapper-multipath-libs packages.

For RHEL6 multipath.conf we typically:

Locate the line that starts "#blacklist {"
Insert comments just above that similar to the following:

# Uncommented blacklist and added DELL PERC and FusionIO to blacklist. Left default comments.

Uncomment the line so it starts just as "blacklist {"

Leave the commented out wwid and devnode lines below that commented then below those insert rules for DELL and Fusion:

device {
vendor “DELL”
product “PERC*”
}
devnode "^fio[a-z]" #Fusion IO

Uncomment the "#}" at end of the blacklist so it closes the open bracket at start.

• Verify user_friendly names section is uncommented and is set to yes:

defaults {
user_friendly_names yes
}

We also Enable multipath daemon:
Run "chkconfig --list multipathd" which should show:
multipathd 0:off 1:off 2:off 3:off 4:off 5:off 6:off

Run "chkconfig multipathd on"

Run "service multipathd start" which should output:
Starting multipathd daemon: [ OK ]

Run "ps -ef |grep multipath" which should show the multipathd now running similar to the following:
root 23225 1 0 17:19 ? 00:00:00 /sbin/multipathd

NOTE:
Every time multipath.conf is changed you must run command to bounce the daemon "service multipathd restart" for it to see the changes.

You can run "multipath -l -v2" to see what multipath thinks is configured.

"Friendly names" are "mpath*" but change on each boot. There is more setup you can do in multipath.conf to give each device a persistent name but that's an advanced topic.