• No se han encontrado resultados

To perform online data migration when EMC PowerPath is present, follow these steps.

NOTE: A maintenance window is required to complete an online data migration with removal of EMC PowerPath (seestep 2).

NOTE: Before using the HP 3PAR Online Import Utility for Oracle RAC=based configurations, it is critical to verify that the cluster registry, voting, and data disks reside on the same source array and are not distributed across multiple arrays.

NOTE: For ASM-based Oracle RAC configurations, the persistent device names that are used by the ASMlib to label ASM disks must be modified. With EMC PowerPath, the devices get a

/dev/emcpower*name, whereas with Linux native device-mapper multipath software, the devices get a/dev/mapper/mpath*name. To ascertain whether the current ASMlib is using the EMC PowerPath-based names, issue the following command:

# oracleasm querydisk -p /dev/emcpower*

EMC PowerPath based-device names are being used if the output is similar to the following example.

Example 16 EMC PowerPath device names

Device "/dev/emcpowera1" is marked an ASM disk with the label "ASM_DISK_*"

1. Zone the host to the destination HP 3PAR StoreServ Storage system to establish communication.

2. Uninstall EMC PowerPath by following these steps:

CAUTION: EMC PowerPath must not be present during the migration process. The native Linux device-mapper-multipath must be managing the paths.

Stop all applications before removing EMC PowerPath. For cluster configurations, also stop the cluster services and the cluster. Applications should be restarted after migration has started.

a. Close all applications on the host.

NOTE: Stop the Oracle RAC database services by issuing the following command on the primary node:

# $ORACLE_HOME/bin/srvctl stop database -d <DB_name> -o immediate

b. Unmount the application file systems, bring offline any raw devices (if configured), and deactivate any volume groups where Logical Volume Manager is in use.

c. For a cluster, stop the cluster services, and then stop the cluster.

NOTE: For Oracle RAC, stop the cluster services and the ASM by performing the following steps:

1) Issue the following command on the primary node: # $GRID_HOME/bin/crsctl stop cluster -all

2) To confirm that the cluster services have stopped, issue the following command on any of the nodes:

# $GRID_HOME/crs_stat -t The expected output is as follows:

CRS-0184: Cannot communicate with CRS Daemon

3) Stop the ASM by issuing the following command on each node: # /etc/init.d/oracleasm stop

d. Uninstall EMC PowerPath from the host, following the EMC PowerPath documentation instructions.

3. Configure the native device-mapper multipath on the host.

• In a cluster environment, update themultipath.conffile on each of the cluster nodes.

NOTE: Edit the/etc/multipath.conffile to map the device WWNs to a persistent device alias. This alias will be used by the ASMlib to label ASM disks for the Oracle RAC application. The/etc/multipath.conffiles should be edited identically on each of the cluster nodes. Make a note of the device type (for instance, ASM data disk, OCR disk, or FRA disk) while naming the disks, and keep them consistent with the way they were configured when EMC PowerPath was enabled. For example:

multipath { wwid 36006016087202e00029a62ca76d3e411 alias <DEVICE_ALIAS>} The value foralias must match the label used by EMC PowerPath for that device. • If the HP 3PAR LUNs are not whitelisted, register the HP 3PAR LUN types with native

device-mapper-multipath by whitelisting HP 3PAR-specific information. (In the/etc/ multipath.conffile, the vendor is3PARdataand the product isVV.) See the HP 3PAR StoreServ product documentation on the Linux host configuration.

NOTE: 3PARdata andVVare case sensitive.

If the HP 3PAR LUNs are whitelisted, start the native device-mapper multipath.

Example 17 Starting the native device-mapper multipath

# /etc/init.d/multipathd restart ok

Stopping multipathd daemon: [ OK ] Starting multipathd daemon: [ OK ]

Example 18 Rescanning HBAs and listing the updated multipath mapping with RHEL 5.x # ls /sys/class/fc_host host2 host3 # echo "1" > /sys/class/fc_host/host2/issue_lip # echo "- - -" > /sys/class/scsi_host/host2/scan # echo "1" > /sys/class/fc_host/host3/issue_lip # echo "- - -" > /sys/class/scsi_host/host3/scan # multipath -ll

mpath13 (360060160bf902a00e03cb1bb3c3fe411) dm-4 DGC,VRAID [size=150G][features=0][hwhandler=0][rw]

\_ round-robin 0 [prio=0][active] \_ 3:0:0:2 sdg 8:96 [active][ready] \_ 2:0:1:2 sdj 8:144 [active][ready]

mpath12 (360060160bf902a00cae2a2a33c3fe411) dm-3 DGC,VRAID [size=200G][features=0][hwhandler=0][rw]

\_ round-robin 0 [prio=0][active] \_ 3:0:0:1 sdf 8:80 [active][ready] \_ 2:0:1:1 sdi 8:128 [active][ready]

mpath11 (360060160bf902a002ac0388f3c3fe411) dm-2 DGC,VRAID [size=200G][features=0][hwhandler=0][rw]

\_ round-robin 0 [prio=0][active] \_ 3:0:0:0 sdb 8:16 [active][ready] \_ 2:0:1:0 sdh 8:112 [active][ready]

4. If EMC PowerPath was uninstalled instep 2, and after multipathing has been configured again on the host, follow these steps:

a. If EMC PowerPath was removed, modify/etc/fstab for new mount points. Base the

new mount points on discovered LUNs, if the migrating LUN alias was not created in the

multipath.conffile. Remount file systems.

IMPORTANT: In Linux, EMC PowerPath devices are presented as/dev/emcpower*

(for example,/dev/emcpowera). Removal of EMC PowerPath in Linux also removes this device type. This changes when the Linux native device-mapper multipath assumes management of these devices. For example, before removal of EMC PowerPath, consider combining/dev/sdband/dev/sdcinto/dev/emcpowera. After removal of EMC PowerPath,/dev/sdband/dev/sdc would be represented by/dev/mpathx. This represents a challenge for customers who use direct device referencing in/etc/fstab

or other custom scripts. HP generally recommends thatfstabmounts be performed using blkid/uuid; however, this is not always employed. In that case, consider mounting /dev/emcpoweraas/var. After removal of EMC PowerPath,/varwould not automatically mount to/dev/mpathx.

b. Where applicable, make the appropriate changes to the Logical Volume Manager configuration.

c. If this is for a cluster configuration, initialize the cluster and start cluster services.

NOTE: For Oracle RAC, edit the /etc/sysconfig/oracleasmfile on both nodes in order to redirect the ASMlib to use the newly defined persistent device names from the

etc/multipath.conffile, which will replace the EMC PowerPath-based /dev/ emcpower*names that were being used to label the ASM disks.

1) In the/etc/sysconfigfile, change the following:

# ORACLEASM_SCANORDER: Matching patterns to order disk scanning ORACLEASM_SCANORDER=""

# ORACLEASM_SCANEXCLUDE: Matching patterns to exclude disks from scan ORACLEASM_SCANEXCLUDE=""

to:

# ORACLEASM_SCANORDER: Matching patterns to order disk scanning ORACLEASM_SCANORDER="dm"

# ORACLEASM_SCANEXCLUDE: Matching patterns to exclude disks from scan ORACLEASM_SCANEXCLUDE="emcpower"

2) Reset theoracleasmservices by issuing the following commands on all cluster nodes, beginning with the primary node:

# /etc/init.d/oracleasm start # oracleasm scandisks

# oracleasm listdisks

3) Verify that all ASM disks are visible from all nodes.

Also verify that the ASM disk labels are now using the appropriate Linux native device-mapper multipath-generated names by issuing the following command: # oracleasm querydisk -p <ASM_label>

The output should be similar to the following example for all configured ASM disks: Device <device_alias> is marked as an ASM disk with the label "ASM_DISK_1"

4) Restart the Oracle RAC cluster and database services by issuing the following command on the primary node:

# $GRID_HOME/bin/crsctl start cluster -all

# $ORACLE_HOME/bin/srvctl start database -d <DB_name>

5) Verify that all cluster services have been restarted by issuing the following commands on all nodes:

# $GRID_HOME/bin/crs_stat —t

# $GRID_HOME/bin/crsctl stat res —t —init

d. Start applications.

5. RHEL 5.x or RHEL 6.x: Using the HP 3PAR Online Import Utility, issue thecreatemigration

command with the-migtype onlineoption.

Optionally, migrations with consistency groups are also supported, using the

-allvolumesincg or-cgvolmap parameters.

NOTE: For Oracle RAC, you must issue thecreatemigrationcommand using the

-allvolumesincg parameter.

Example 19 Usingcreatemigrationwith-migtype onlinefor RHEL 5.x or RHEL 6.x

# createmigration -sourceuid xxxxxxxxxxxxxxxx -srchost "X7DBU-02-IG" -destcpg testcpg -destprov full

-migtype online -persona “RHEL_5_6”

# SUCCESS: Migration job submitted successfully. Please check status/details using showmigration command. Migration id: 1411673030253

For Oracle RAC: Using the HP 3PAR Online Import Utility, issue thecreatemigration

command with the-migtype onlineand-allvolumesincgparameters.

Example 20 Usingcreatemigrationwith -migtype onlinefor Oracle RAC

# createmigration -sourceuid xxxxxxxxxxxxxxxx -srchost "x8dtuf-01" -allvolumesincg -migtype online -destcpg

testcpg -destprov full -persona "RHEL_5_6"

This step admits LUNs presented to the EMC Storage system and exports them to the HP 3PAR host created on the EMC Storage system.

NOTE: Thecreatemigration command may take several minutes to complete.

TIP: Make a note of the migration ID, as it will be used in commands to track migration progress.

6. Issue theshowmigrationcommand to verify that the data migration task preparation has completed successfully. This may take some time. Upon successful creation of the

createmigrationtask, theSTATUScolumn in theshowmigrationcommand output will indicatepreparationcomplete(100%). When this status is indicated, continue to the next step.

Example 21 Issuing theshowmigrationcommand

# showmigration

MIGRATIONID TYPE SOURCE_NAME DESTINATION_NAME START_TIME

END_TIME STATUS(PROGRESS)(MESSAGE)

1411673030253 online SYMMETRIX+000194901212 3par_7200_DCA_01 Thu Sep 25 15:

23:50 EDT 2014 -NA- preparationcomplete(100%)(-NA-)

7. Using the HP 3PAR Management Console, verify that the host whose LUNs are under migration has paths to as many HP 3PAR controller nodes as are zoned in the SAN (seeFigure 2 (page 47)). This number of paths is listed in theNode ID(s)column. The path details are in theHost Detailspane.

Figure 2 HP 3PAR Management Console—Host Details

8. Update path configuration on the host by rescanning all HBAs and issuing themultipath -llcommand to verify the newly discovered paths. Multipath now recognizes extra paths to the HP 3PAR StoreServ Storage, but all paths are still managed as source array devices.

Example 22 Rescanning HBAs and listing the updated multipath mapping with RHEL 5.x # ls /sys/class/fc_host host2 host3 # echo "1" > /sys/class/fc_host/host2/issue_lip # echo "- - -" > /sys/class/scsi_host/host2/scan # echo "1" > /sys/class/fc_host/host3/issue_lip # echo "- - -" > /sys/class/scsi_host/host3/scan # multipath -ll

mpath13 (360060160bf902a00e03cb1bb3c3fe411) dm-4 DGC,VRAID [size=150G][features=0][hwhandler=0][rw] \_ round-robin 0 [prio=0][active] \_ 2:0:0:2 sde 8:64 [active][ready] \_ 3:0:0:2 sdg 8:96 [active][ready] \_ 2:0:1:2 sdj 8:144 [active][ready] \_ 3:0:1:2 sdm 8:192 [active][ready]

mpath12 (360060160bf902a00cae2a2a33c3fe411) dm-3 DGC,VRAID [size=200G][features=0][hwhandler=0][rw] \_ round-robin 0 [prio=0][active] \_ 3:0:0:1 sdf 8:80 [active][ready] \_ 2:0:0:1 sdd 8:48 [active][ready] \_ 2:0:1:1 sdi 8:128 [active][ready] \_ 3:0:1:1 sdl 8:176 [active][ready]

mpath11 (360060160bf902a002ac0388f3c3fe411) dm-2 DGC,VRAID [size=200G][features=0][hwhandler=0][rw] \_ round-robin 0 [prio=0][active] \_ 3:0:0:0 sdb 8:16 [active][ready] \_ 2:0:0:0 sdc 8:32 [active][ready] \_ 2:0:1:0 sdh 8:112 [active][ready] \_ 3:0:1:0 sdk 8:160 [active][ready]

9. Delete the migrating LUN paths pertaining to the source array. For details about identifying and deleting LUN paths, see“Appendix B: Identifying and Deleting Source Array LUN Paths” (page 105).

10. Unzone the host from the source array.

Example 23 Output after unzoning the host from the source array with RHEL 5.x

# multipath —ll

mpath13 (360060160bf902a00e03cb1bb3c3fe411) dm-4 3PARdata,VV [size=150G][features=0][hwhandler=0][rw]

\_ round-robin 0 [prio=0][active] \_ 2:0:0:2 sde 8:64 [active][ready] \_ 3:0:1:2 sdm 8:192 [active][ready]

mpath12 (360060160bf902a00cae2a2a33c3fe411) dm-3 3PARdata,VV [size=200G][features=0][hwhandler=0][rw]

\_ round-robin 0 [prio=0][active] \_ 2:0:0:1 sdd 8:48 [active][ready] \_ 3:0:1:1 sdl 8:176 [active][ready]

mpath11 (360060160bf902a002ac0388f3c3fe411) dm-2 3PARdata,VV [size=200G][features=0][hwhandler=0][rw]

\_ round-robin 0 [prio=0][active] \_ 2:0:0:0 sdc 8:32 [active][ready] \_ 3:0:1:0 sdk 8:160 [active][ready]

11. Start the migration by using the HP 3PAR Online Import Utilitystartmigrationcommand.

Example 24 Starting migration by usingstartmigration # startmigration -migrationid 1411673030253 # SUCCESS: Data transfer started successfully.

NOTE: The-migrationid(in the example above,1411673030253) will have been assigned by thecreatemigrationcommand.

TIP: TheSTATUScolumn in the showmigrationcommand output will indicatesuccess

when all volumes have been migrated successfully.

Example 25showmigrationcommand showing successful migration

# showmigration -migrationid 1411673030253

MIGRATION_ID TYPE SOURCE_NAME DESTINATION_NAME

START_TIME END_TIME STATUS(PROGRESS)(MESSAGE)

1411673030253 online CLARiiON+APM00 3par_7200_DCB_01

Fri Apr 04 16:38:24 EDT 2014 -NA- success(-NA-)(-NA-)

Documento similar