Problem Unmapping Lun 0 with MPXIO on Sun Storage[TM] 351x FC Arrays

Asset ID:	1-72-1007843.1
Update Date:	2010-12-20
Keywords:

Solution Type Problem Resolution Sure

Solution 1007843.1 : Problem Unmapping Lun 0 with MPXIO on Sun Storage[TM] 351x FC Arrays

Applies to:

Sun Storage 3511 SATA Array
Sun Storage Traffic Manager Software - Version: 3.0 and later [Release: 3.0 and later]
Sun Storage 3510 FC Array - Version: Not Applicable and later [Release: N/A and later]
All Platforms

Symptoms

{SYMPTOM}

Symptoms

A logical drives (LD) is mapped as LUN 0 on multiple channels.

After unmapping the LD on one channel the Sun Storage[TM] 351x FC Array will present an ESI device as
LUN 0. Mpxio will pair off this new ESI device with the remaining LUN's of the LD on the other channels.

When the host looses access to the LD , format displays it as:

    1. c8t600C0FF00000000000660D1AE0FCA401d0 
/scsi_vhci/ssd@g600c0ff00000000000660d1ae0fca401

Detailed Example of the Symptom:

3 Sun Storage[TM] 351x FC Array partitions are mapped on channel 0 and 4. Mpxio is enabled:

# cfgadm -al -o show_SCSI_LUN
Ap_Id                          Type         Receptacle   Occupant     Condition
c2                             fc-private   connected    configured   unknown
c2::256000c0ffc0660d,0         disk         connected    configured   unknown
c2::256000c0ffc0660d,1         disk         connected    configured   unknown
c2::256000c0ffc0660d,2         disk         connected    configured   unknown
c3                             fc           connected    unconfigured unknown
c4                             fc-private   connected    unconfigured unknown
c5                             fc-private   connected    configured   unknown
c5::216000c0ffc0660d,0         disk         connected    configured   unknown
c5::216000c0ffc0660d,1         disk         connected    configured   unknown
c5::216000c0ffc0660d,2         disk         connected    configured   unknown

LUN 0 is mapped on channel 0 and 4 of the 351x array.
The disk is seen from the host (by the format command) and labeled using type 0
(Auto configure). Luns can be used without any problem.

Then all LUNs on channel 4 are unmapped. The SE351x presents an ESI device to the host
on channel 4.

# cfgadm -al -o show_SCSI_LUN
Ap_Id                          Type         Receptacle   Occupant     Condition
c2                             fc-private   connected    configured   unknown
c2::256000c0ffc0660d,0         ESI          connected    configured   unknown
c3                             fc           connected    unconfigured unknown
c4                             fc-private   connected    unconfigured unknown
c5                             fc-private   connected    configured   unknown
c5::216000c0ffc0660d,0         disk         connected    configured   unknown
c5::216000c0ffc0660d,1         disk         connected    configured   unknown
c5::216000c0ffc0660d,2         disk         connected    configured   unknown

In format, the LUN 0 is now shown and cannot be labeled.
LUN 0 is unusable while luns 1 and 2 are working fine.

# format
Searching for disks...done
AVAILABLE DISK SELECTIONS:
0. c0t0d0 
/pci@1f,4000/scsi@3/sd@0,0
1. c8t600C0FF00000000000660D1AE0FCA401d0 
/scsi_vhci/ssd@g600c0ff00000000000660d1ae0fca401
2. c8t600C0FF00000000000660D1AE0FCA403d0 
/scsi_vhci/ssd@g600c0ff00000000000660d1ae0fca403
3. c8t600C0FF00000000000660D0431530F00d0 
format> type
AVAILABLE DRIVE TYPES:
0. Auto configure
........
22. other
Specify disk type (enter its number): 0
Auto configure failed
format> quit

From luxadm display it looks like mpxio paired off the ESI device on channel 4,
and the LUN 0 on channel 0. It still reports both channels as online, even though
there is only one channel left.

# luxadm display /dev/rdsk/c8t600C0FF00000000000660D1AE0FCA401d0s2
DEVICE PROPERTIES for: /dev/rdsk/c8t600C0FF00000000000660D1AE0FCA401d0s2
Vendor:               SUN
Product ID:           StorEdge 3511
Revision:             413C
Serial Num:           00660D000000
Unformatted capacity: 500000.000 MBytes
Write Cache:          Enabled
Read Cache:           Enabled
Minimum prefetch:   0x0
Maximum prefetch:   0xffff
Device Type:          SES device
Path(s):

 /dev/rdsk/c8t600C0FF00000000000660D1AE0FCA401d0s2
/devices/scsi_vhci/ssd@g600c0ff00000000000660d1ae0fca401:c,raw
Controller           /devices/pci@1f,4000/SUNW,qlc@2/fp@0,0
Device Address              256000c0ffc0660d,0
Host controller port WWN    210000e08b0aafaf
Class                       primary
State                       ONLINE
Controller           /devices/pci@1f,4000/pci@4/SUNW,qlc@5/fp@0,0
Device Address              216000c0ffc0660d,0
Host controller port WWN    210000e08b098b2e
Class                       primary
State                       ONLINE

For LUN 1, luxadm is reporting one channel as online, the other as offline.

# luxadm display /dev/rdsk/c8t600C0FF00000000000660D1AE0FCA403d0s2
DEVICE PROPERTIES for disk: /dev/rdsk/c8t600C0FF00000000000660D1AE0FCA403d0s2
Vendor:               SUN
Product ID:           StorEdge 3511
Revision:             413C
Serial Num:           00660D1AE0FC
Unformatted capacity: 768951.000 MBytes
Write Cache:          Enabled
Read Cache:           Enabled
Minimum prefetch:   0x0
Maximum prefetch:   0xffff
Device Type:          Disk device
Path(s):

 /dev/rdsk/c8t600C0FF00000000000660D1AE0FCA403d0s2
/devices/scsi_vhci/ssd@g600c0ff00000000000660d1ae0fca403:c,raw
Controller           /devices/pci@1f,4000/SUNW,qlc@2/fp@0,0
Device Address              256000c0ffc0660d,1
Host controller port WWN    210000e08b0aafaf
Class                       primary
State                       OFFLINE
Controller           /devices/pci@1f,4000/pci@4/SUNW,qlc@5/fp@0,0
Device Address              216000c0ffc0660d,1
Host controller port WWN    210000e08b098b2e
Class                       primary
State                       ONLINE

Changes

{CHANGE}

Cause

{CAUSE}

Solution

Resolution

To recover from this situation a reboot of the host is required.

There are two ways available to anticipate the occurence of this problem.

1. For LUN 0 devotees:

Create small dummy partitions for both SE351x controllers and map them as
LUN 0 on all channels. Do not use LUN masking for those LUN 0, as all hosts
need to see it. Do not touch those LUN's any more. Leaving those LUN 0's
available all the time prevents the presentation of ESI devices as lun 0.

2. For all others:

Do not use LUN 0 for the mapping of LD's at all. The SE351x will present ESI
devices as LUN 0 on all channels in the case when nothing is mapped to LUN 0. A LUN 0
is required to comply with the SCSI standard. The problem only occurs
when unmapping a LUN 0. Therefore not mapping an LD to LUN 0 makes it
difficult to trigger the problem.

minnow, SE3511, mpxio, drive type unknown, lun 0

Change History
Date: 2010-12-20
User Name: susan.copeland@oracle.com
Action: Address Comments (add 3510)

Attachments

This solution has no attachment