Skip to main content
NetApp Knowledge Base

Are file clone splits generated by OpenStack Cinder

  1. Last updated
  2. Save as PDF
Views:
Visibility:
Public
Votes:
0
Category:
trident-openshift
Specialty:
openstack
Last Updated:

Applies to

  • ONTAP v9.xx
  • AFF-A150 (and similar AFF/ASA platforms)
  • OpenStack Cinder using NetApp NFS driver
  • (FlexVol volumes) Environments with image caching and large-volume cloning (e.g., 100GB Windows images)
  • Customers observing clone splits during volume/instance creation and deletion
  • Performance concerns

Answer

  • This is not true.
  • Clone splits are background ONTAP processes that make these clones fully independent by copying shared blocks.
  • When Cinder creates a new volume from an image or clones a volume, ONTAP uses space-efficient file clones (SIS clones) that share data blocks with the parent.
  • Splits are automatically queued by ONTAP under certain conditions, such as:
    • The parent snapshot or file needs to be deleted or modified.
    • The FlexVol’s space utilization crosses certain thresholds.
    • The clone relationship would block other operations.
  • Cinder itself does not initiate clone splits (neither during create nor delete).
  • The ONTAP storage system manages splits as needed to maintain data integrity and independence.
  • If performance is a concern during the ONTAP split operations, refer to the following below:
  • Throttle Clone Split Load (Recommended):
    • Use the ONTAP CLI or REST API to adjust the maximum split load per node, which controls how many splits can run concurrently and limits their I/O impact.
    • volume file clone split load modify -node <node> -max-split-load <value>
  • Increase OnDemand Job Limit (Optional for Large-Scale Testing):
    • For environments with high split concurrency, increase the ONTAP bootarg ondemand.job_limit (default is 50, can be raised to 500).
    • This allows more concurrent split jobs, reducing queueing time for splits. Requires node takeover/giveback to take effect.
  • Monitor Split Activity:
    • Regularly monitor split queue depth and load using:
      • volume file clone split status 
      • volume file clone split load show
  • Most environments with warm data and modern ONTAP hardware experience minimal performance impact unless splits queue up faster than they can complete.
  • Consider FlexGroup Volumes for Scale:
    • At large scale, FlexGroup volumes distribute clone/split operations, reducing bottlenecks compared to FlexVol.
  • Image Caching:
    • Ensure Cinder image caching is enabled to maximize fast clones and minimize full-copy operations.
  • If split operations are causing noticeable performance issues during peak hours, temporarily throttle the split load as above.
  • Alternatively, schedule large-scale clone/delete operations during off-peak periods to minimize impact.

Additional Information

Why does flex volume clone split take a long time?

OpenStack VM Creation Performance Issues with NetApp ONTAP Cinder Driver and Cross-Pool Cloning

NetApp provides no representations or warranties regarding the accuracy or reliability or serviceability of any information or recommendations provided in this publication or with respect to any results that may be obtained by the use of the information or observance of any recommendations provided herein. The information in this document is distributed AS IS and the use of this information or the implementation of any recommendations or techniques herein is a customer's responsibility and depends on the customer's ability to evaluate and integrate them into the customer's operational environment. This document and the information contained herein may be used solely in connection with the NetApp products discussed in this document.