EMC CLARiiON

EMC CLARiiON CX4 model 480 provides high-capacity networked storage for demanding online transaction processing (OLTP) workloads and large-scale email environments. With CLARiiON CX4 model 480, you can scale seamlessly up to 471 TB of storage capacity and consolidate two times the workload in one array compared to any other storage provider.

Use proven CLARiiON five 9s availability and innovative technologies—fully automated storage tiering (FAST), FAST Cache, Flash drives, compression, 64-bit operating system, and multicore processors. Enjoy UltraFlex technology for dual-protocol, online-expandable connectivity options and future technology integration.(Source www.emc.com)

EMC CX4 is an asymmetric active-active array.

HP-UX: For 11.31 it supports ALUA and for 11.23 it works in active/passive mode with the native PV link failover s/w. For 11.31 the initiator type should be set as non-auto trespass and for 11.23 auto trespass should be used with failover mode as 4 and 0 respectively. With Powerpath 5.1 or earlier the failover mode must be set as 1 (even for Powerpath 5.1 or later in 11.23 )and for Powerpath 5.1.1 or later it can be set as 4 or 1 in 11.31.

Solaris: Supports ALUA with Veritas 5.0 MP3 RP1. Till 5.0 MP3 it works in active/passive mode. In case of ALUA IO goes through all paths (both optimal & non-optimal) whereas in active/passive IO goes only through active paths (if we have 2 optimal and 2 non-optimal then IO will go through 2 optimal paths).

#vxdmpadm iostat show all

cpu usage = 1388228us per cpu memory = 32768b

OPERATIONS BLOCKS AVG TIME(ms)

PATHNAME READS WRITES READS WRITES READS WRITES

c1t0d0s2 350142 1656907 14407947 14796647 0.06 0.94

c3t50001FE1502582EDd4s2 167496 108193 78145063 3179047 0.01 0.23

c3t50001FE1502582E9d4s2 140694 347855 27247192 11273827 0.05 0.53

c2t50001FE1502582ECd4s2 38411 21604 13068509 405469 0.02 0.68

c2t50001FE1502582E8d4s2 50314 62517 11027584 1561229 0.06 2.22

The above command shows IO for all paths. Additional commands used for display and configuring FC & disks

#vxdisk list

#vxdisk path

#devfsadm

#fcinfo hba-port

#fcinfo remote-port -s -p PWWN

#cfgadm

#cfgadm -al

#cfgadm -c configure c#

Whenever DMP is used with any storage, the 1st thing that should be done is to check the ASL (Array Support Library) of Veritas for the specific storage.

#vxddladm listsupport

#vxddladm listsupport libename=library-name

With Powerpath till 5.0 failover must be set as 1 and with 5.1 or later it can be set as 4 for ALUA behavior.

AIX: With AIX primarly used multipathing s/w are Powerpath & Veritas DMP. Initiator type should be set as Clariion Open and the failover mode can be set as below.

Powerpath 5.3 or earlier, it must be set as 3.

Powerpath 5.3 SP1 and above it can be set as 3 for active/passive or 4 for ALUA behavior.

PowerPath 5.3 is the minimum version for AIX 6.1 running on a VIO server.

PowerPath 5.1 is the minimum version for AIX 6.1 running on a standard image.

AIX also requires EMC Clariion ODM to work properly.If running AIX 5.3 then EMC Clarrion 5.3.0.3 ODM should be installed and if it's 6.1 then it should be 5.3.0.4 (minimum required is 5.3.0.3).

#lslpp -al |grep -i emc

EMC.CLARiiON.aix.rte 5.3.0.4 COMMITTED EMC CLARiiON AIX Support

EMC.CLARiiON.fcp.rte 5.3.0.4 COMMITTED EMC CLARiiON FCP Support

EMCpower.base 5.3.0.0 COMMITTED PowerPath Base Driver and

EMCpower.consistency_grp 5.3.0.0 COMMITTED PowerPath Consistency Group

EMCpower.encryption 5.3.0.0 COMMITTED PowerPath Encryption with RSA

EMCpower.migration_enabler

EMCpower.mpx 5.3.0.0 COMMITTED PowerPath Multi_Pathing

EMC.CLARiiON.aix.rte 5.3.0.4 COMMITTED EMC CLARiiON AIX Support

EMC.CLARiiON.fcp.rte 5.3.0.4 COMMITTED EMC CLARiiON FCP Support

LINUX: With LINUX primarily used multipath s/w are powerpath & DM-MPIO. When using DM-MPIO no additional rpm should be installed. The installed kernel device mapper rpm very much compatible with CX4, just need to configure it (/etc/multipath.conf , /etc/modprobe.conf, /etc/lvm/lvmconf).

Novell SLES:

Novell SLES 10.1 (SP1) or above: failovermode may be set to 1 for active/passive behavior, or

Novell SLES 10.1 (SP1) or above: failovermode may be set to 4 for ALUA behavior

Novell SLES 10.0 (no SP) or below: failovermode MUST be set to 1

RedHat RHEL 4.x:

RHEL 4.6 or above: failovermode may be set to 1 for active/passive behavior, or

RHEL 4.6 or above: failovermode may be set to 4 for ALUA behavior

RHEL 4.5 or below: failovermode MUST be set to 1

RedHat RHEL 5.x:

RHEL 5.1 or above: failovermode may be set to 1 for active/passive behavior, or

RHEL 5.1 or above: failovermode may be set to 4 for ALUA behavior

RHEL 5.0: failovermode MUST be set to 1

ESX: ESX support both Powerpath & VMware Native multipath s/w.

Vmware Native multipath:

ESX 3.x the failover must be set as 1.

ESX 4.x the failover can be set as 1 for active/passive or 4 of ALUA behavior.

Powerpath:

Failover can be set as 1 for active/passive or 4 for ALUA behavior.

In ESX when set as active/passive then IO goes through only 1 path irrespective of whether there are 2 active and 2 standby paths if the default path policy (Most Recently Used is set). If the path policy is changed to Round Robin then IO goes through all of the optimal paths.The recommended path policy for 3.x is fixed policy and for 4.x it's is Round Robin (Most Recently Used path policy can also be used). If ALUA is set as failover then all the paths will show as active but there would be optimal and non-optimal paths.

Some of the primarily used Powerpath commands

#powermt check

#powermt save

#powermt version

#powermt restore

#powermt display

#powermt display dev=all

#powermt display dev=hdiskpower#

#emcgreg -list

#emcpreg -add key [key---]

Some of the primarily used Navicli & Naviserverutil commands

#navicli -h SPA-IP getcache

#navicli -h SPA-IP getdisk 2_0_1

#navicli -h SPA_IP getlun LUN# -default -owner

#echo 1 |/opt/Navisphere/bin/naviserverutil

When PV link is used in 11.23 admins need to make sure that the 1st PV mentioned in Volume group is an active path. Together with navicli, naviserverutil and ioscan commands the active path can be determined

#echo 1 |/opt/Navisphere/bin/naviserverutil

c4t1d0 LUN 259 10.19.171.247 B

c8t1d0 LUN 259 10.19.171.247 B

#/opt/Navisphere/bin/navicli -h 10.19.171.246 getlun 259 -default -owner

Default Owner: SP B

Current owner: SP B

#ioscan -kfnCdisk |grep -i t1d0

/dev/dsk/c3t1d0 /dev/rdsk/c3t1d0

/dev/dsk/c4t1d0 /dev/rdsk/c4t1d0

/dev/dsk/c7t1d0 /dev/rdsk/c7t1d0

/dev/dsk/c8t1d0 /dev/rdsk/c8t1d0

In the above example cxt1d0 is a disk from CX and it has 2 active and 2 passive paths. From the naviserverutil we came to know that c4 and c8 are coming from SP B so c3 and c7 are from SP A.

The above navicli command showed that the current and default owner of the LUN is SP B. So we need to select either c4 or c8 as the 1st device in Volume group.

If the above navicli command had given default owner as SP B and current owner as SP A then the naviserverutil command would have given c3 and c7 with SP A in it's output. So we need to keep in mind that naviserverutil give us the current LUN owner and not the default owner.Otherwise the LUN would get trespassed. Also we need to disable the proactive polling for the alternate links.

#pvchange -p n /dev/dsk/c3t1d0

If the proactive polling is not disabled then there would be massive trespassing of the LUN which may adversely the performance of the CX.