XtremIO X1: All Software Modules restart after a volume is removed

           

   Article Number:     538338                                   Article Version: 3     Article Type:    Break Fix 
   

 


Product:

 

XtremIO Family,XtremIO X1

 

Issue:

 

 

   

      The following scenario describes the failure flow on a Volume or a Consistency Group with one or more volumes in it:   

   
         
  •         One or more Snapshot Sets are created from the Consistency Group OR one or more Snapshots ate created from the Volume     
  •      
  •         The Snapshot-Set/snapshot is removed (either via an RSET or CG removal in the RecoverPoint environment or a Snapshot-Set/Snapshot removal in XtremIO with no RecoverPoint)     
  •      
  •         The Volume or Consistency Group are removed.     
  •      
  •         All software modules in the cluster will restart - a short, temporary performance impact may be seen as the software modules restart.     
  •    
   
      This issue is limited to XtremIO version 4.0.27-1   
                                                             

 

 

Cause:

 

 

   

      When a Consistency Group is removed, all of its Snapshot Sets are updated to point to an invalid Short ID (SNID). This is a fail-safe measure to prevent RecoverPoint target data inconsistency when SNID numbers are reused. If at this point, a volume(s) included in that Snapshot Set is removed, the software modules will restart due to failure when trying to correlate between the internal volume ID and the SNID as the SNID is invalid until such time as the relevant Snapshot 'merge' processes are complete.   

                                                             

 

 

Resolution:

 

 

   

      To prevent this issue from occurring:   

   
         
  1.         Remove all of the CG's/Volume's dependent SnapshotSets/Snapshots (On a Recoverpoint connected to the XtremIO cluster the removal is done by deleting the CG/RSET)     
  2.      
  3.         Allow enough time  for all the snapshot merge processes to complete (depends on many factors so we recommend waiting 2-3 days).     
  4.      
  5.         On the XtremIO cluster unmap the parent CG/volume from the host(s).      
  6.      
  7.         Remove the parent CG/Volume on the XtremIO cluster.     
  8.