Move Cluster to New Site Procedures

Cluster and Data Center Move Procedures




Cluster Move Procedure to New Site - Cluster is Re-IPed


This procedure is used when a cluster needs to move to a new data center and failover, failback and re-ip of a cluster is required at the new data center.




Steps need to follow while moving one cluster to a new location:


  1. Shares or exports exist on active cluster (writeable data) and on the read only cluster (cluster B)

  2. Failover one policy  or access zone at a time to the target cluster (Cluster B)

    1. Expected Result: Clients and applications are successfully accessing data from the new active cluster (cluster B)

      1. Completes without error

      2. Shares and exports now exist on new active cluster (cluster B) on the new read only cluster (cluster A)

  3. Disable all Eyeglass jobs

  4. Move source (cluster A) to new site

  5. Change  IP address on Cluster A

  6. Bring source (cluster A) back online

  7. Change source (cluster A) IP address from Eyeglass inventory view using the below procedure:

http://documentation.superna.net/eyeglass-isilon-edition/igls-administration/edit-isilon-ip-address-in-eyeglass

    1. Expected Result: Eyeglass Configuration Replication Jobs are running without error

  1. Failback one policy at a time or access zone to the source (cluster A)

    1. Expected Result:

      1. Clients and applications are successfully accessing data from the new active cluster (cluster A)

      2. Eyeglass configuration replication Jobs on new active cluster (cluster A) may be in error and can be corrected with the procedure below

  2. Eyeglass jobs should be all be enabled at this step.  Confirm this is the case.

  3. Reset Eyeglass to pick up IP address changes for Jobs on the new active cluster (cluster A)

    1. SSH to Eyeglass appliance using admin: sudo -s enter (must use root)  then use admin password (default password: 3y3gl4ss)

    2. Run command <igls admin rediscover>

    3. Once reset completes, go to the chrome browser and refresh the browser and login with the credentials

    4. Login to jobs window and validate that all jobs are in the correct states, eyeglass will restore all job status but should be double checked,

      1. Correct any if jobs in user disabled state or not in DFS state that should be DFS enabled.

      2. IMPORTANT: You must enable DFS mode on an job that did not rediscover and shows in the DFS job section of jobs window before enabling the Job to prevent creation of active shares on target cluster. Enable DFS mode first, then enable.

    5. Using Bulk Actions select a policy in jobs window and use Run now.

      1. Using Running jobs tab to watch and verify successful (check mark no red X) on the configuration sync

    6. If using Access Zone failover, verify that failover readiness job is enabled.

      1. Using bulk actions force run the readiness job wait until it completes (running jobs window)

    7. Check DR Dashboard to verify it shows a good DR status

  4. Done








Cluster Move Procedure to New Site - Cluster IP Unchanged


Steps need to follow while moving one cluster to a new location: using same network setting (same cluster IP address )


Before moving the cluster to new location, configuration replication jobs in Eyeglass should be working properly and configuration synced to target cluster.


  1. Disable configuration replication jobs in Eyeglass before turning off the cluster being moved.

Note: For Failover scenario, the failover away from the cluster that is going to be moved should be completed successfully

- config replication jobs that synced the configuration from source to target in disable mode

- Jobs on the active cluster are green.


Login to Eyeglass  →  click on Jobs icon in desktop →   Job Definitions → check the box to select jobs in Configuration Replication:Share, Export, Alias replication and/or Configuration Replication:Access Zone replication   (if applicable) and/or

Configuration Replication:DFS mode  (if applicable) and/or

Configuration Replication:Snapshot Schedules (if applicable) and/or

Failover: Runbook Robot section (if applicable)  → Select bulk a action → Enable/Disable to disable the jobs




  1. Then Turn off the moved cluster.


  1. After moving the cluster to the new site and powered up. Check the connectivity between the cluster and Eyeglass appliance, hint: you may ping the cluster  IP from Eyeglass.


  1. Enable Configuration Replication:Share,Export, Alias replication jobs in Eyeglass.   


Login to Eyeglass  →  click on Jobs icon in desktop →   Job Definitions → check the box to select jobs in Configuration Replication:Share, Export, Alias replication  section → Select bulk a action → Enable/Disable to enable the jobs.

Note: Enable any Snapshot Schedules, Configuration Replication:Access Zone replication, Failover: Runbook Robot  jobs were enabled before cluster move.


  1. Run the configuration for one replication cycle to verify configuration replication jobs working properly.  

click on Jobs icon in desktop →   Job Definitions → check the box to select jobs in Configuration Replication:Share, Export, Alias replication  section → Select bulk a action → Run now




Steps need to follow while moving one cluster to a new location: using same network setting (Different cluster  IP address )


Before moving the custer to new location, configuration replication jobs in Eyeglass should be working properly and configuration synced to target cluster.


  1. Disable configuration replication jobs in Eyeglass before turn the source cluster off. For Failover scenario. Failover should be completed successfully - config replication jobs that synced the the configuration in disable mode.  


Login to Eyeglass  →  click on Jobs icon in desktop →   Job Definitions → check the box to select jobs in configuration replication:shares, exports, alias replication  section → Select bulk a action → Enable/Disable to disable the jobs

Notes:

DFS jobs should be disabled; if any, before turn off the cluster.

From Eyeglass 1.6.3, Snapshot Schedules and Access Zone replication should be disabled.


  1. Turn off the Source cluster.


  1. Reset Eyeglass appliance.


Reset Eyeglass  to re-add the cluster (different IP address will be added to eyeglass for the moved cluster) .

SSH to Eyeglass appliance using admin: sudo -s enter (must use root)  then use admin password (default password: 3y3gl4ss)

  1. cd /opt/superna/sbin

  2. ./reset.sh

  3. Once reset completes, login to chrome browser and refresh the browser and login with the credentials.

  4. Now, add both of clusters using Management subnet SSIP.


  1. After moving the cluster to the new site and powered up. Check the connectivity between the cluster and Eyeglass appliance, hint: you may ping the eyeglass IP.



  1. Enable Configuration Replication:Share,Export, Alias replication jobs in Eyeglass.   


Login to Eyeglass  →  click on Jobs icon in desktop →   Job Definitions → check the box to select jobs in Configuration Replication:Share, Export, Alias replication  section → Select bulk a action → Enable/Disable to enable the jobs.

Note: Enable any Snapshot Schedules and Access Zone replication jobs were enabled before cluster move.



  1. Run the configuration for one replication cycle to verify configuration replication jobs working properly.  

click on Jobs icon in desktop →   Job Definitions → check the box to select jobs in Configuration Replication:Share, Export, Alias replication  section → Select bulk a action → Run now





Useful links : How to - Edit Isilon IP Address in Eyeglass


http://documentation.superna.net/eyeglass-isilon-edition/igls-administration/edit-isilon-ip-address-in-eyeglass