Why Did the Used Storage Shrink After Scale-out?
Possible Causes
If you do not run VACUUM to clear and reclaim the storage space before the scale-out, the data deleted from GaussDB(DWS) may not free up the occupied disk space.
During the scale-out, the system redistributes the data because the service data volume on the original nodes is significantly larger than that on the newly added nodes. When the redistribution starts, the system automatically performs VACUUM to free up the storage space. This causes a big drop in capacity.
Handling Procedure
You are advised to periodically clear and reclaim the storage space by running VACUUM FULL to prevent data expansion.
If the used storage space is still large after you run VACUUM FULL, analyze whether the existing cluster flavor meets service requirements. If no, scale out the cluster.
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.
