IBM FlashSystem

IBM FlashSystem

Find answers and share expertise on IBM FlashSystem

 View Only
  • 1.  FlashGrid setup error

    Posted 10 days ago

    Setting up a new FlashGrid and having some issues.  The error is a bit ambiguous.  I'm getting the error: 

    Message ID

    CMMVC6035E
    Error text Replication configuration is blocked because the partner system already has an object of the same type and name as one being replicated. The type and name of the object are shown in table Details of remote command error

    The detail of the error is here (Volume name obfuscated)

    Partner system    <Partnersystemname>
    Error Object    partition <PartitionName>
    Command type    Internal command run as part of creating a host
    Other object ID in command    0
    Message ID    CMMVC6035E

    This customer has two sites approx 7 miles apart and has dedicated dark fiber between sites, response times are solid.  For a subset of their vmware hosts, they have volumes that are presented from Flashsystem A at Site A to Host A, but they also have volumes from FlashSystem B at Site B to Host A over the dark fiber. My assumption is that the Partition setup is erroring out because it sees the host definition for Host A at Site A, but on the Site B system.  Thinking it was a naming conflict I changed the name of the Site B Hosts, but the error still persists.  They are coming from a Hyperswap configuration which we will be doing an image-mode migration to this once the structure is setup and functional. 

    My assumption is that the host object itself is causing the problem regardless of naming.  Is this a restriction of FlashGrid and Policy Based HA?  I can't have the same host definition on both ends of the HA Pairing?



    ------------------------------
    Shaun Anderson
    ------------------------------


  • 2.  RE: FlashGrid setup error

    Posted 7 days ago

    Hi Shaun,

    it is about the Host and the according host-port definitions.

    PBHA does not support a multiple usage of a WWPN. Even if you did a rename of the existing host object, the WWPNs you want to use within PBHA are already confired at the partner system.

    BTW: If you want to map replicated and non replicated volumes to a host, the current solution is to either use additional fc cards or NPIV WWPNs (if the host OS does support this, e.g. PowerVM) or do a sparation by protocol, e.g. NVMe vs. SCSI.

    Regards, Uwe



    ------------------------------
    Uwe Schreiber
    System Engineer / Solution Architect
    SVA System Vertrieb Alexander
    Wiesbaden
    ------------------------------



  • 3.  RE: FlashGrid setup error

    Posted 5 days ago

    This is a rough drawing of our intended direction.  The distance will easily support PBHA.  So I am confused as to which level I should create objects.  For the 'CH' side, a given volume group has volumes mapped to the ESXi clusters at both sites today.  They have done some Storage DRS work in the past so while we can quibble with their decisions, this is what it is.  If I create a host definition at the CH-7300 System level, then I can't create the same host definition inside of the partition it seems.  Likewise, if I create both site host definitions for a particular partion (for instance going CH -> CC) then I'm not able to create that host definition for the volumes that will stay on the CC side nor for the volumes that will be in a volume group presented to hosts on both sides, as well as replicating CC -> CH.

    I destroyed the partition, removed all host mappings, and attempted to sync just the vg with volumes.  They do not exist on the target side and I'm still getting the 4100 error with sense data that isn't terribly helpful. 

    In it's simplest permutation I can grasp what we're trying to accomplish with FlashGrid.  However I'm wondering if it's gameday ready for more advanced configurations yet.  I see the Redbook was published on 7/30, the day after I got home from the customer site and I have read through that  but it's not giving examples that help a user understand anything beyond basic A to B replication/HA. And in it it states:

    This setup is to replace a Hyperswap configuration if that helps with context.  

    My ask:

    • Is what I'm defining clear?
    • Is it possible to have the same host definition exist within a partition and natively within the array?
    • Is a hyperswap-esque configuration not really an option with flashgrid if the above are restrictions?

    Appreciate the input.



    ------------------------------
    Shaun Anderson
    ------------------------------



  • 4.  RE: FlashGrid setup error

    Posted 7 days ago

    Hello Shaun,

    The error is saying that partition ID 0 on the partner system has the same name as the partition it's currently trying to make HA with that system. Partition names (as well as hosts, volume groups, volumes and host clusters) need to be unique within a system. Rename the partition on either system to something else and mark the event as fixed and it will try again.



    ------------------------------
    Chris Bulmer

    Software architect, IBM Storage Virtualize
    ------------------------------



  • 5.  RE: FlashGrid setup error

    Posted 5 days ago

    I don't think that's it, I think we're firmly in bug territory....OR...there is some backend cleanup that takes time to clear?  I wiped the partition that was created as ID 0.  Currently I have ID 1 and 2 that are both working as PBHA  (no hosts, just volumes) One going CH -> CCC, the other CCC -> CH. Detail below:

    CH View
    id:name:preferred_management_system_name:active_management_system_name:replication_policy_id:replication_policy_name:location1_system_name:location1_status:lm_name:location2_status:host_count:volume_group_count:ha_status:link_status:desired_location_system_name:migration_status:draft:draft_volume_group_count:draf
    1:partJoCoCH-AIX:JOCO_CH_FS7300:JOCO_CH_FS7300:2:ha-policy-2:JOCO_CH_FS7300:healthy:JOCO_CCC_FS7300:healthy:0:1:established:synchronized:::no:0:0
    2:partJoCoCCC-VMWare:JOCO_CCC_FS7300:JOCO_CCC_FS7300:3:ha-policy-3:JOCO_CCC_FS7300:healthy:JOCO_CH_FS7300:healthy:0:1:established:synchronized:::no:0:0

    CCC View
    id:name:preferred_management_system_name:active_management_system_name:replication_policy_id:replication_policy_name:location1_system_name:location1_status:lm_name:location2_status:host_count:volume_group_count:ha_status:link_status:desired_location_system_name:migration_status:draft:draft_volume_group_count:draf
    1:partJoCoCH-AIX:JOCO_CH_FS7300:JOCO_CH_FS7300:2:ha-policy-2:JOCO_CH_FS7300:healthy:JOCO_CCC_FS7300:healthy:0:1:established:synchronized:::no:0:0
    2:partJoCoCCC-VMWare:JOCO_CCC_FS7300:JOCO_CCC_FS7300:3:ha-policy-3:JOCO_CCC_FS7300:healthy:JOCO_CH_FS7300:healthy:0:1:established:synchronized:::no:0:0

    CH View
    vgJIMS-cccVMWARE    7      partJoCoCCC-VMWare
    vgJIMS-chVMWARE        30    
    vgJIMSchAIX        7    partJoCoCH-AIX

    CCC View
    vgJIMS-cccVMWARE    7    partJoCoCCC-VMWare
    vgJIMSchAIX        7    partJoCoCH-AIX

    I recreated the partition giving it the vgJIMS-chVMWARE Volume Group.  The replication establishes, and shortly thereafter gives the same error and drops the replication into suspended state:

    Event ID
    051010
    Event ID Text
    A remote command failed during replicated storage partition configuration.
    Sequence Number
    258
    Object Type
    partition
    Object ID
    0
    Object Name
    partJoCoCH-VMWare

    So the partition didn't exist, the VG only existed on the source side.  It appears that it is creating the object on the destination and then erroring out saying it already exists.  

    I can get a ticket opened on this issue, but any help on my more procedural questions would be appreciated.

    Thanks y'all.



    ------------------------------
    Shaun Anderson
    ------------------------------