CX4 cross connect between SPA and SPB

May i know the output for :

powermt display dev=all 

Thanks

C:\>powermt display dev=all

Pseudo name=harddisk1

Logical device ID=60060160F7802800FAF0B65DB74EDF11 [LUN 0]

state=alive; policy=CLAROpt; priority=0; queued-IOs=0;

Owner: default=SP A, current=SP A       Array failover mode: 1

==============================================================================

--------------- Host ---------------   - Stor -   -- I/O Path --  -- Stats ---

###  HW Path               I/O Paths    Interf.   Mode    State   Q-IOs Errors

==============================================================================

   7 port7\path0\tgt0\lun0    c7t0d0    SP B0     active  alive       0      0

   8 port8\path0\tgt0\lun0    c8t0d0    SP A0     active  alive       0      0

Pseudo name=harddisk2

Logical device ID=60060160F7802800329BFA64B74EDF11 [LUN 1]

state=alive; policy=CLAROpt; priority=0; queued-IOs=0;

Owner: default=SP A, current=SP A       Array failover mode: 1

==============================================================================

--------------- Host ---------------   - Stor -   -- I/O Path --  -- Stats ---

###  HW Path               I/O Paths    Interf.   Mode    State   Q-IOs Errors

==============================================================================

   7 port7\path0\tgt0\lun1    c7t0d1    SP B0     active  alive       0      0

   8 port8\path0\tgt0\lun1    c8t0d1    SP A0     active  alive       0      0

Pseudo name=harddisk3

Logical device ID=60060160F780280008724D91B84EDF11 [LUN 2]

state=alive; policy=CLAROpt; priority=0; queued-IOs=0;

Owner: default=SP A, current=SP A       Array failover mode: 1

==============================================================================

--------------- Host ---------------   - Stor -   -- I/O Path --  -- Stats ---

###  HW Path               I/O Paths    Interf.   Mode    State   Q-IOs Errors

==============================================================================

   7 port7\path0\tgt0\lun2    c7t0d2    SP B0     active  alive       0      0

   8 port8\path0\tgt0\lun2    c8t0d2    SP A0     active  alive       0      0

Pseudo name=harddisk4

Logical device ID=60060160F7802800D8A0AE9FB84EDF11 [LUN 3]

state=alive; policy=CLAROpt; priority=0; queued-IOs=0;

Owner: default=SP A, current=SP A       Array failover mode: 1

==============================================================================

--------------- Host ---------------   - Stor -   -- I/O Path --  -- Stats ---

###  HW Path               I/O Paths    Interf.   Mode    State   Q-IOs Errors

==============================================================================

   7 port7\path0\tgt0\lun3    c7t0d3    SP B0     active  alive       0      0

   8 port8\path0\tgt0\lun3    c8t0d3    SP A0     active  alive       0      0

Any reason why you have allocated the production LUNs to one SP and backup LUNs to the other ?

Also, how you concluded Powerpath is not not working ?

So the Idea will be SPA A0 and SPB B0 able to access all the LUN in SPA

Peter,

you can look at path usage by running "powermt display dev=all every=5", as well as "powermt display latency dev=all every=5". Another thing that i noticed that you only have 2 paths per device. It is considered best practice to zone each HBA to both storage processors for loadbalacnig/failover purposes.

HBA1 - SPA0, SPB1

HBA2 - SPB0, SPA1

..or something of that nature.

dear Dynamox,

i'm not sure about the powerpath setup it was my first attempt. is my setup consider correct ? actually i'm trying to get EMC engineer to check and do it for me , but non of them replied to me yet.

my problem was when i did the sqlio stress test, the IO took very long time and the disk queue was skyhigh.

your setup from connectivity perspective can be improved, let's put it that way . It will add more paths to the SP that owns the LUN, giving you more queues in the OS.

your queues and high response time could be due to the LUN you are testing against not being able to handle the IO you are pushing to it. Do you have Navisphere Analyzer installed on this array ?

Hi Dynamox,

meaning the configuration all correct now ? yes i plan to contact emc to add a san switch to add more path to my storage.

meanwhile how do i use the analyzer to test on the IO and performance ?

Dear Dynamox ,

how can i add more path to the SP ?

we have navisphere analyzer installed, how we can use it to analyze the situation?

Dear Rupal,

how can i change the mode to 4 , is it using powermt ?

thanks.

dear Dynamox,

i've successfully changed the failover mode to 4 , but when i running the stress test, the SP B0 in power path still showing no activities, does it means it didnt share to load of the IO yet ?

Pseudo name=harddisk4

Logical device ID=60060160F7802800D8A0AE9FB84EDF11 [LUN 3]

state=alive; policy=CLAROpt; priority=0; queued-IOs=0;

Owner: default=SP A, current=SP A       Array failover mode: 4

==============================================================================

--------------- Host ---------------   - Stor -   -- I/O Path --  -- Stats ---

###  HW Path               I/O Paths    Interf.   Mode    State   Q-IOs Errors

==============================================================================

   7 port7\path0\tgt0\lun3    c7t0d3    SP B0     active  alive       0      0

   8 port8\path0\tgt0\lun3    c8t0d3    SP A0     active  alive       0      0

Power path console

SP A0

c8t0d0          optimal          Alive          Active          585          2

c8t0d1          optimal          Alive          Active          0          2

c8t0d2          optimal          Alive          Active          763          25

c8t0d3          optimal          Alive          Active          0          2

SP B0

c7t0d0          optimal          Alive          Active          0          0

c7t0d1          optimal          Alive          Active          0          0

c7t0d2          optimal          Alive          Active          0          0

c7t0d3          optimal          Alive          Active          0          0

oh, this is direct connect right now ? Then it makes sense that you can't add more paths (unless you have extra HBAs that is)

In Navisphere/Univsphere click on the Help option, there should be a very good section on how to use Analyzer.

that is correct because on Clariion/VNX a LUN is own by one storage process so you will only see i/o to either SPA or SPB. What you can do is to add more paths to give you more "lanes" to drive i/o through.

you need to physically connect additional ports from Clariion to your switches. You want each switch to have connections from SPA and from SPB. To change failover mode you will need to go into Connectivity Status and edit host registration there, make sure to reboot the host imediatly after changing the value.

yes. . i think i will contact my EMC partner to get more in depth details for my current setup.

thanks.