Forum Discussion

abhagat's avatar
abhagat
Novice I
4 years ago

Hello Team, Need help in Pure Metro storage cluster configuration

Hello Team, Need help in Pure Metro storage cluster configuration. https://support.purestorage.com/Solutions/VMware_Platform_Guide/User_Guides_for_VMware_Solutions/ActiveCluster_with_VMware_User_Guide/vSphere_Metro_Storage_Cluster_With_ActiveCluster%3A_Overview_and_Introduction As per this document there are two ways to configure it Uniform and non-uniform. Uniform configuration indicates Local and Remote paths to be configured as AO (Active Optimized), but during storage failure scenario Local storage takes 17 to 20 seconds to time out. During that time Storage array sends a PDL to ESXi host and host removes the devices. Due to which the vSphere HA configured to react on PDL restarts the VMs and that causes a outage for the customer. Which is not expected. From ESXI host there is no way to Change PDL behavior. Now my question is there anything that we are missing in this configuration, as during my talk with Consultants they are very sure 17 to 20 seconds delay is expected in this setup and during that time though the paths are available devices returned D:0x2 0x5 0x25 which is PDL. Looking for the guidance to overcome this situation. If anybody here faced such situation during the implementation

virtualization

2 Replies

  • mmoran's avatar
    mmoran
    Puritan
    4 months ago

    Make sure to check the Host Personality and Set Personality: ESXi. Hope this helps!

    The ESXi personality does the following as of Purity 5.1.0:

    • It makes the FlashArray issue a Permanent Device Loss SCSI sense response to ESXi when a pod goes offline due to a mediator loss. If this is not set, no response is sent, and vSphere HA does not detect the failure properly. It will not restart VMs running on the failed hosts.
    • ESXi uses peripheral LUN IDs instead of flat LUN IDs—this changes how ESXi views any LUN ID on the FlashArray above 255. Since ESXi does not properly interpret flat LUN IDs, it sees LUN ID higher than 255 to be 16,383 higher than it should be (256 is seen as 16,639), which is outside of the supported range of ESXi. Setting the ESXi personality on the FlashArray for a given host switches the FlashArray LUN methodology to peripheral, allowing ESXi to see LUN IDs higher than 255.

     

    https://support.purestorage.com/bundle/m_user_guides_for_vmware_solutions/page/Solutions/VMware_Platform_Guide/User_Guides_for_VMware_Solutions/FlashArray_VMFS_RDM_VMware_Best_Practices_User_Guide/library/common_content/t_setting_the_flasharray_esxi_host_personality.html

  • jase's avatar
    jase
    Puritan
    4 months ago

    When the local FlashArray is offline, those paths are down, but because the devices are present from both FlashArray's alternate paths are available. 

    If your testing results in a PDL when bringing the local FlashArray offline, it doesn't appear that it is in a uniformed configuration. 

    Can you confirm that paths are available to both FlashArrays? When both FlashArrays are available, the local FlashArray (to your local hosts) should have paths listed as Active (I/O) and the paths to the remote FlashArray (to your local hosts) should have paths listed as Active.