Sun Storage 6130 Array
 
Sun Storage Traffic Manager Software
 

GCS>Sun Microsystems>Storage - Disk>Modular Disk - 6xxx Arrays
 
GCS>Sun Microsystems>Storage Software>Storage Resource Management
 

This document describes how to migrate volumes on a Sun StorEdge[TM] 6130 array,
from the preferred controller to the alternate controller, using both the
Element Manager and sscs(1M) interfaces. Included, will be how to affect that
change for the data host, using either Veritas Dynamic Multi-Pathing(DMP), or
Sun StorEdge[TM] Traffic Manager(STMS) software.

This document applies to solutions(host, switch, array), in which the data path
and array health, are otherwise optimal.

Note: Element Manager and sscs utilities are provided as part of the Sun StorEdge[TM]
6130 Array Software installation for the management host in your solution.
See the StorEdge 6130 Getting Started Guide for more information.

Part A: Before Beginning

Since this array is an Asymmetrical type of array, the data paths to the
controllers, are either primary or secondary. Once the volumes,
that are to have the preferred controller change are identified, disable the
primary path to that device, on the Data Host.

For STMS software, or STMS with DMP

Sun StorEdge Traffic Manager(STMS), makes it really easy to identify the logical
device path for a volume. Look at the Volume Details screen in Element
Manager, or sscs list -a <array_name> volume <volume_name>. The
World-Wide Number(WWN) value in those outputs, will give you the target string
for the logical device path in Solaris OS.

Example:

 # ./sscs list -a pts-toi-6130 volume vol0-32k-r1 | grep WWN: | grep -v Initiator
WWN:              60:0A:0B:80:00:17:0E:1C:00:00:0C:1A:42:6C:E0:30
#

If the ":" is removed, it is a device path, and can be matched using
luxadm probe.

Example:

 # luxadm probe
No Network Array enclosures found in /dev/es
#

Found Fibre Channel device(s):
Node WWN:200400a0b8170e1c  Device Type:Disk device
Logical Path:/dev/rdsk/c5t600A0B8000170F660000196E4270CF85d0s2

Now you have a device path. To view the individual path status',
run luxadm display <logical path>.

Example:

 # luxadm display /dev/rdsk/c5t600A0B8000170F660000196E4270CF85d0s2
DEVICE PROPERTIES for disk: /dev/rdsk/c5t600A0B8000170F660000196E4270CF85d0s2
Vendor:               SUN
Product ID:           CSM100_R_FC
Revision:             0610
Serial Num:           1T43528121
Unformatted capacity: 1024.000 MBytes
Write Cache:          Enabled
Read Cache:           Enabled
Minimum prefetch:   0x0
Maximum prefetch:   0x0
Device Type:          Disk device
Path(s):

  /dev/rdsk/c5t600A0B8000170F660000196E4270CF85d0s2
/devices/scsi_vhci/ssd@g600a0b8000170f660000196e4270cf85:c,raw
Controller           /devices/pci@8,600000/SUNW,qlc@1/fp@0,0
Device Address              200500a0b8170e1d,1
Host controller port WWN    210000e08b076cb5
Class                       primary
State                       ONLINE
Controller           /devices/pci@8,600000/SUNW,qlc@1,1/fp@0,0
Device Address              200400a0b8170e1d,1
Host controller port WWN    210100e08b276cb5
Class                       secondary
State                       STANDBY

Notice that it shows that the primary path is ONLINE, as opposed to STANDBY.
This is a healthy set of connections. If one or more primary paths show as
STANDBY or some other state, stop and investigate, before proceeding. There
could be something wrong with the device paths.

Finally, you can use luxadm failover secondary <logical raw device path>,
to change the primary paths to a STANDBY state.

Example:

 # luxadm failover secondary /dev/rdsk/c5t600A0B8000170F660000196E4270CF85d0s2
#

Check luxadm display <path> for the change.

Example:

 # luxadm display /dev/rdsk/c5t600A0B8000170F660000196E4270CF85d0s2
DEVICE PROPERTIES for disk: /dev/rdsk/c5t600A0B8000170F660000196E4270CF85d0s2
Vendor:               SUN
Product ID:           CSM100_R_FC
Revision:             0610
Serial Num:           1T43528230
Unformatted capacity: 1024.000 MBytes
Write Cache:          Enabled
Read Cache:           Enabled
Minimum prefetch:   0x0
Maximum prefetch:   0x0
Device Type:          Disk device
Path(s):

   /dev/rdsk/c5t600A0B8000170F660000196E4270CF85d0s2
/devices/scsi_vhci/ssd@g600a0b8000170f660000196e4270cf85:c,raw
Controller           /devices/pci@8,600000/SUNW,qlc@1/fp@0,0
Device Address              200500a0b8170e1d,1
Host controller port WWN    210000e08b076cb5
Class                       primary
State                       STANDBY
Controller           /devices/pci@8,600000/SUNW,qlc@1,1/fp@0,0
Device Address              200400a0b8170e1d,1
Host controller port WWN    210100e08b276cb5
Class                       secondary
State                       ONLINE

For Veritas Dynamic Multi-Pathing(DMP) ONLY

1) Find the primary and secondary paths

   vxdmpadm getsubpaths dmpnodename=<device>

  Example

   # vxdmpadm getsubpaths dmpnodename=T30_0

NAME        STATE[A]   PATH-TYPE[M] CTLR-NAME  ENCLR-TYPE   ENCLR-NAME    ATTRS
================================================================================
c2t20020F23000101B5d2s2 ENABLED    PRIMARY      c2         6130           61300              -
c1t20020F230001020Ad2s2 ENABLED(A) SECONDARY    c1         6130           61300              -

Note: If subpaths show an STMS path, recognizable by a string that is longer
than those in the example above in scsi target value(cNt<value>dNsN), then you
are using STMS with DMP, so follow the above directions for STMS.

2) Disable the Primary path

 vxdmpadm disable ctlr=<CTLR-NAME>

 Example:

    #vxdmpadm disable ctlr=c2
#

Part B: Management Host Procedure

To change the Preferred controller, using Sun StorEdge[TM] Configuration Service

1) List the volume details for the volume in question

  USAGE:

   sscs list -a <array_name> volume <volume_name> | grep Controller

   example

   # ./sscs list -a 6130 volume myvol | grep Controller
Controller:                      B
Preferred Controller:            A
#

NOTE: Be sure that the Preferred Controller does not match the Controller ''value. This indicates the the failover commands issued in part A were performed correctly. ''Matching values for a volume after performing the steps in part A, indicate that either the array is not healthy, or the steps in part A were not performed correctly. If they do match, it is advisable to stop, and check the host /var/adm/messages file for any path problems.

2) Change the preferred controller value

  USAGE:

  sscs modify -a <array_name> -c <A | B> volume <volume_name>

  Example:

   # ./sscs modify -a 6130 -c B volume myvol
#

3) Check the preferred controller value

  USAGE:

   sscs list -a <array_name> volume <volume_name> | grep Controller

  Example 

   # ./sscs list -a 6130 volume myvol | grep Controller
Controller:                      B
Preferred Controller:            B
#

To Change the Preferred Controller Using Element Manager

1) From the Array Summary screen, click on your array name

2) Under the Storage -> Logical -> Volumes screen click on your volume
name to get to the Volume Details screen.

3) Scroll down to the Owning Controller drop-down menu, and change the
value to the opposite controller of that shown in the Preferred Controller
field.

4) Click OK

NOTE: If this fails to change the Preferred Controller, or if the two fields
match before you change the value of the Owning Controller field, check
the health of your array controllers, network connections, volumes, and vdisks.
''This is due to the manually initiated failover performed in the previous
sections.''

5) Review the Preferred Controller and Owning Controller fields, they should
now be of the selection you made in step 3, and they should be exactly the same
value.

Part C: Data Host Procedure

If Using SSTM or SSTM with DMP

By changing the Preferred controller, the device path for the primary and
secondary paths to the array are changed. The array throws no change
notification, hence SSTM will not pick up the change in how it needs to handle
the paths.

'Switched Fabric or SAN(Storage Area Network) Configurations:'

To reconfigure the device paths:

   cfgadm -c configure <controller number> 

This will reset the path dedications of primary and secondary, and
place them in the ONLINE and STANDBY states as appropriate. Use this for
all HBA paths, that can see the LUN which has been changed. The
/var/adm/messages file, will have evidence of the reconfiguration, if you look
for messages from the scsi_vhci driver.

'In a Direct Attached Storage(DAS) Configurations:'

SSTM requires a system reboot to make the change. As always, the shutdown(1M)
or init(1M) commands should be used.

If Using DMP Only

At this time, you can only perform a system boot to correct the paths. Neither
the use of the vxdctl command, or any of the variants of the vxconfigd commands
will make the appropriate change in which path is primary and
secondary.

If the directions above were followed, and the primary path was disabled using
vxdmpadm, the system boot will enable the path automatically.

BugID 6204434 for details on why SSTM requires a reboot for DAS configurations

This solution has no attachment