Reformatting and Resetting a CSN Swarm cluster

To reformat and reset a CSN-managed Swarm cluster, you will follow this process:

1. Delete the Elasticsearch index.
2. Reformat the Swarm volumes.
3. Reset the CSN's logs and other data.
4. Reset the CSN configuration backups.

 Details:

1. Delete the Swarm Elasticsearch index (if a Search Feed had been configured).
2. Reformat the disks
   1. Attach a keyboard and monitor to a Swarm storage server. 
   2. Select System Control > Stop Storage Processes.
   3. Select Disk Volumes > All and press F to Format, and accept the confirmations.
   4. Repeat with each server, leaving the storage processes off (so that there is no replication to the newly formatted disks)
   5. After all storage server disks have been formatted and storage processes have been turned off, on all storage servers, select System Control > Reboot.
   6. Watch to ensure that they PXEboot.
3. Reset the CSN's logs and other cluster-specific data
   1. Log into the CSN via SSH.
   2. Remove any unwanted logs in /var/log/caringo/.  
      • Relocate those to retain for historical reference (especially cr.log (if any), castor.log, and scspproxy.log), but you will want fresh logs for the fresh cluster. 
   3. On older CSNs that included the Content Router, delete any data in /u01/ 
   4. On older CSNs, remove any rrd files in /var/opt/caringo/StorageReporter/data/
4. Reset the CSN backups
   
   1. Create a new backup manifest UUID by removing the UUID in this file: 
      /var/opt/caringo/csn/backup/manifest/.bootstrap 
   2. Remove references to the old backups by deleting the savesets files in:
      /var/opt/caringo/csn/backup/manifest/
   3. Reboot your CSN.  
   4. Log into the CSN UI.  
   5. Make a manual backup: Cluster Services > Backup and Restore > Create Backup.
   6. Note the new backup manifest UUID.

See also: How to reset a CSN development environment