Forum Discussion

mikebounds's avatar
mikebounds
Level 6
10 years ago

Can vxdisk -p list show information from both controllers

I have 2 paths to disks - example:

 # vxdmpadm -v getportids dmpnodename=hitachi_usp-v0_0c19
NAME              ENCLR-NAME     ARRAY-PORT-ID    pWWN    SFG_ID
================================================================
c3tWWN45d5s2      hitachi_usp-v0       5F         WWN45   5
c1tWWN55d5s2      hitachi_usp-v0       6F         WWN55   6

 

But when I run vxdisk -p list it gives me information about one of the controllers:

# vxdisk -x PORT_SERIAL_NO -x ARRAY_PORT_PWWN -x ARRAY_CTLR_ID  -p list hitachi_usp-v0_0c19
DEVICE                PORT_SERIAL_NO   ARRAY_PORT_PWWN   ARRAY_CTLR_ID
hitachi_usp-v0_0c19   5F               WWN45             5

 

Even I specify the actual disk path for the "6F" controller, I still get information about 5F controller

# vxdisk -x PORT_SERIAL_NO -x ARRAY_PORT_PWWN -x ARRAY_CTLR_ID  -p list c1tWWN55d5s2
DEVICE                PORT_SERIAL_NO   ARRAY_PORT_PWWN   ARRAY_CTLR_ID
hitachi_usp-v0_0c19   5F               WWN45             5

 

Can I get vxdisk -p list to give me information on the other controller?

Mike

 

  • Hi Tony, I think it is fine to have a limitation that "vxdisk -p list" does not show information about paths, but this needs to be all or nothing - I don't think it is right to show information about an arbrituary path when you enquire about a disk, so vxdisk -p list shouldn't show disk port is 5F when it is actually dual pathed via 5F and 6F.  In partilcular in Starflyfly's example , vxdisk -p list shows 2 LUNs are using different ports, but in fact they are using the same ports - so vxdisk -p list is giving misleading information.

    Mike

  • Similar results here. Looks like 'vxdisk -p list' reports array info for the device associated with the DMP node name only, and not the other active path(s).

    # vxdisk path
    SUBPATH                     DANAME               DMNAME       GROUP        STATE
    ...
    sdau                        hitachi_hus_vm1_0900 hitachi_hus_vm0_0900 oradg1f      ENABLED
    sdga                        hitachi_hus_vm1_0900 hitachi_hus_vm0_0900 oradg1f      ENABLED
    ...


    # vxdmpadm getsubpaths all
    NAME         STATE[A]   PATH-TYPE[M] DMPNODENAME  ENCLR-NAME   CTLR   ATTRS
    ==================================================================
    ...
    sdau         ENABLED(A)   -          hitachi_hus_vm1_0900 hitachi_hus_vm1 c0        -
    sdga         ENABLED(A)   -          hitachi_hus_vm1_0900 hitachi_hus_vm1 c1        -
    ...


    # vxdisk -x PORT_SERIAL_NO -x ARRAY_PORT_PWWN -x ARRAY_CTLR_ID  -p list
    DEVICE               PORT_SERIAL_NO ARRAY_PORT_PWWN         ARRAY_CTLR_ID
    ...     
    hitachi_hus_vm1_0900 3A             50:06:0e:80:13:2c:1d:20 3           
    ...

     

  • Seems yes.

     

    c2t500601603022FB3Fd12s2 ENABLED(A) PRIMARY      emc_clariion0_32 emc_clariion0 c2        -
    c2t500601683022FB3Fd12s2 ENABLED    SECONDARY    emc_clariion0_32 emc_clariion0 c2        -
    c3t500601603022FB3Fd12s2 ENABLED(A) PRIMARY      emc_clariion0_32 emc_clariion0 c3        -
    c3t500601683022FB3Fd12s2 ENABLED    SECONDARY    emc_clariion0_32 emc_clariion0 c3        -

     

    #  vxdisk -x PORT_SERIAL_NO -x ARRAY_PORT_PWWN -x ARRAY_CTLR_ID  -p list

    DEVICE             PORT_SERIAL_NO ARRAY_PORT_PWWN         ARRAY_CTLR_ID 
    ...
    emc_clariion0_32   B0             50:06:01:68:30:22:fb:3f B             
    emc_clariion0_33   A0             50:06:01:60:30:22:fb:3f A             
    emc_clariion0_34   A0             50:06:01:60:30:22:fb:3f A             
    emc_clariion0_35   B0             50:06:01:68:30:22:fb:3f B             
    emc_clariion0_36   A0             50:06:01:60:30:22:fb:3f A

    ...            

  • Starflyfly, 

    From your output you have 2 WWNs - one with 68 and one with 60, so your vxdmpadm getsubpaths extract is saying you have a singe LUN called emc_clariion0_32 which has 4 paths and these paths are on port WWN 68 and 60 (and I guessing the switches that c2 and c3 are connected to are joined which is why you see 4 paths and not the normal 2).

    So then vxdisk -p list is only showing port A0 for emc_clariion0_32 and port B0 for emc_clariion0_33, which would suggest the 2 LUNs are on different ports, but I suspect both LUNS are on BOTH A0 and B0 ports, which means the vxdisk -p list is even more misleading than mine as at least mine shows all LUNS as been on port 5F and none on port 6F.

    Coud you confirm this by providing output from:

    vxdmpadm -v getportids dmpnodename=emc_clariion0_32
    
    
    vxdmpadm -v getportids dmpnodename=emc_clariion0_33

     

    Thanks

    Mike

  • Hi, Mike

     

    vxdmpadm -v getportids dmpnodename=emc_clariion0_32
    NAME                 ENCLR-NAME     ARRAY-PORT-ID                   pWWN            SFG_ID
    ============================================================================================
    c3t500601683022FB3Fd12s2 emc_clariion0        B0               50:06:01:68:30:22:fb:3f 1        
    c3t500601603022FB3Fd12s2 emc_clariion0        A0               50:06:01:60:30:22:fb:3f 2        
    c2t500601603022FB3Fd12s2 emc_clariion0        A0               50:06:01:60:30:22:fb:3f 9        
    c2t500601683022FB3Fd12s2 emc_clariion0        B0               50:06:01:68:30:22:fb:3f 10       
    # vxdmpadm -v getportids dmpnodename=emc_clariion0_33
    NAME                 ENCLR-NAME     ARRAY-PORT-ID                   pWWN            SFG_ID
    ============================================================================================
    c3t500601683022FB3Fd13s2 emc_clariion0        B0               50:06:01:68:30:22:fb:3f 1        
    c3t500601603022FB3Fd13s2 emc_clariion0        A0               50:06:01:60:30:22:fb:3f 2        
    c2t500601603022FB3Fd13s2 emc_clariion0        A0               50:06:01:60:30:22:fb:3f 9        
    c2t500601683022FB3Fd13s2 emc_clariion0        B0               50:06:01:68:30:22:fb:3f 10   

  • vxdisk path:

     

    c3t500601683022FB3Fd12s2    emc_clariion0_32     emc_clariion0_32 testdg       ENABLED
    c3t500601603022FB3Fd12s2    emc_clariion0_32     emc_clariion0_32 testdg       ENABLED
    c2t500601603022FB3Fd12s2    emc_clariion0_32     emc_clariion0_32 testdg       ENABLED
    c2t500601683022FB3Fd12s2    emc_clariion0_32     emc_clariion0_32 testdg       ENABLED
    c3t500601603022FB3Fd13s2    emc_clariion0_33     -            -            ENABLED
    c3t500601683022FB3Fd13s2    emc_clariion0_33     -            -            ENABLED
    c2t500601683022FB3Fd13s2    emc_clariion0_33     -            -            ENABLED
    c2t500601603022FB3Fd13s2    emc_clariion0_33     -            -            ENABLED

     

     

    # vxdmpadm getsubpaths all
    ..
    c2t500601603022FB3Fd12s2 ENABLED(A) PRIMARY      emc_clariion0_32 emc_clariion0 c2        -
    c2t500601683022FB3Fd12s2 ENABLED    SECONDARY    emc_clariion0_32 emc_clariion0 c2        -
    c3t500601603022FB3Fd12s2 ENABLED(A) PRIMARY      emc_clariion0_32 emc_clariion0 c3        -
    c3t500601683022FB3Fd12s2 ENABLED    SECONDARY    emc_clariion0_32 emc_clariion0 c3        -
    c2t500601603022FB3Fd13s2 ENABLED(A) PRIMARY      emc_clariion0_33 emc_clariion0 c2        -
    c2t500601683022FB3Fd13s2 ENABLED    SECONDARY    emc_clariion0_33 emc_clariion0 c2        -
    c3t500601603022FB3Fd13s2 ENABLED(A) PRIMARY      emc_clariion0_33 emc_clariion0 c3        -
    c3t500601683022FB3Fd13s2 ENABLED    SECONDARY    emc_clariion0_33 emc_clariion0 c3        -

  • So I was correct - both LUNS are on BOTH A0 and B0 ports, but vxdisk -p list says port A0 for emc_clariion0_32 and port B0 for emc_clariion0_33, so this is incorrect, or very misleading at best.

    My thoughts are vxdisk -p list should either:

    1. Show information about all controllers
    2. Not show any information about controllers
    3. Show information on controller requested

    But it should not show information about an arbrituaty controller as the consequence of this in Starflyfly's configuration is that vxdisk -p output implies emc_clariion0_32 and emc_clariion0_33 are connected differently (on different ports), but in fact they are the same.

    Even if I give vxdisk -p the path of disk on controller 6F in my example, it still shows as being connected on port 5F (last output on my opening post).

    So it would seem you can't get info about all controllers or even a specific controller and you can only get information on the controller  vxdisk -p happens to choose.

    Mike

  • Hi Mike,

    This appears to be a current limitation with this command and operation

    best regards

    tony

  • Hi Tony, I think it is fine to have a limitation that "vxdisk -p list" does not show information about paths, but this needs to be all or nothing - I don't think it is right to show information about an arbrituary path when you enquire about a disk, so vxdisk -p list shouldn't show disk port is 5F when it is actually dual pathed via 5F and 6F.  In partilcular in Starflyfly's example , vxdisk -p list shows 2 LUNs are using different ports, but in fact they are using the same ports - so vxdisk -p list is giving misleading information.

    Mike