Forum Discussion

ch1's avatar
ch1
Day Hiker III
1 month ago
Solved

Active Cluster for File

Hello everyone :-)

I'm new to Active Cluster for File and I'm looking for some documentation or introductory material.

I'd like to understand:

  • How Active Cluster for File works
  • How it is configured via the management interface
  • General setup, requirements, and administration tasks

If anyone can point me to official documentation, knowledge base articles, or deployment guides, I would really appreciate it.

Thanks!

7 Replies

  • Hi Ch1, 

    You can test  ActiveCluster for File failover by initiating a failover via software or by pulling power to both controllers on the active array.  From a software perspective, the purenetwork eth remediate command can be used for this.   You can also find the remediate option in the GUI as well.   The below offers an example of testing failover of realm called realm-test-0,   who's VIF is active on Array1, and triggers failover to Array2.

    # Check the current VIF availability (should be on Array1)
    $> purenetwork eth list --vif-availability
    
    # Update the preferred array list to the partner array (this stages the remediation)
    $> purenetwork eth setattr --preferred-arraylist <Array2> --remediation-mode manual realm-test-0::server1::vif0001# Verify both sites are online$> purerealm list realm-test-0
    
    # Execute the manual remediation (failover)
    $> purenetwork eth remediate realm-test-0::server1::vif0001
    
    # Check the VIF availability status again (should now be on Array2)
    $> purenetwork eth list --vif-availability
    
    # Verify the realm status post-failover
    $> purerealm list realm-test-0

    You can also test ACF failover by disconnecting power to both of the controllers on the array that has the active VIF on it as well,  to test unplanned failovers as well.   

    Please let me know if you have additional questions.

    Cheers, 

    Quinn

    • ch1's avatar
      ch1
      Day Hiker III

      Hello Quinn,

      sorry about the late response. I have massive issues replying to post and on dms. If I press reply on posts or in dms, nothing happes. But after 1 week it seems like I can reply again. 

      Unfortunately the failover with the commands didn't work.
      The second command:
      "purenetwork eth setattr --preferred-arraylist <Array2> --remediation-mode manual realm-test-0::server1::vif0001"

      Gives following error message:
      "Error on VIF_name: Specified network interface does not exist."


      I am using the VIF which is given trough the first command
      "purenetwork eth list --vif-availability"


      Fortunately we could succesfully test the failover with the remediate option in the GUI. The failover works :-)


      Thank you, your post was very helpful! 

      Kind Regards
      Kevin 

      • thomasroeder's avatar
        thomasroeder
        Day Hiker II

        Hello,

        I had to add the interfaces for both arrays—as shown in the following example—to both FAs in order for the remediation to work.

         

  • Ludes's avatar
    Ludes
    Community Manager

    Hey Kevin - thanks for reaching out here in the Community.  There's a recently published ActiveCluster for File quick start guide:

    https://support.everpuredata.com/r/activecluster/activecluster-file-quick-start-guide

    Also a solution overview and best practices linked here:

    https://support.everpuredata.com/r/activecluster/activecluster-for-file-synchronous-file-resiliency-requirements-and-best-practices

    qsummers​ is a great contact if you want to DM him.  Let us know your experience herebas you get going and reach out with any other questions,.  

  • Hello,

    I have another question. I am currently implementing Active Cluster for FILE, but regarding the server interface configuration—specifically for the Realm server—I need to add the sub-interfaces for each array, correct? For example, I add `ct0.eth1` for Array 1; then, since `array1:ct0.eth1` is automatically replicated, I should edit it to add the `ct0.eth1` on Array 2, and likewise add the `Array2:ct0.eth1` interface on Array 1, right? To test this, I make the change using Bulk Edit, select the "Remediate" option, and I am able to ping the interface IP. Is this the correct procedure?

    I didn't find any mention of this "Remediate" procedure in the articles, only in the GUI itself.

    Thank you.

     

  • ch1's avatar
    ch1
    Day Hiker III

    ​ Ludes​ thank you, for your reply! The second link with the "best practices" unfortunaly doesnt work. 

    ​ 
    epark​ Thank you also, for you reply! The video was very helpful!

    The Information was very helpful. We were able to create a realm, a server, polcies, NFS file share and we succesflly mount the NFS share.

    What we are trying to do right now: Testing a failover

    We testet a failover by switching off the NFS and replication Ports on the switch. Unfortunately, the failover did not work. Is there antoher way to trigger a failover?

    In the video, at 18:43 a failover is performed, but I am not sure how it is initiated.  
    https://www.everpuredata.com/demos/platform/what-is-the-pure-storage-platform/introducing-activecluster-for-file-on-flasharray/6390644453112.html

    I have found an option called "realm connections". Maybe this option has to be configured before a failvoer can happen?

    Does anyone have an Idea how we can test an failover before we can take this system live?

    Thank