Versions (relevant - OpenSearch/Dashboard/Server OS/Browser):
AWS managed OpenSearch 2.17
Describe the issue:
An automated snapshot is created each hour.
After the indexation of a big vector indice (> 410 gb, 256 dimension, 12 primaries, 0 replica), the creation of an automated snapshot is running, the status of the new vector index is always STARTED during more than 5 hours, so I can’t do open or search action for this index. And I met several times AutomatedSnapshotFailure, I should add a OpenSearch node.
Why the creation of an automated snapshot is slowly for big vector index? How to revole the problem of AutomatedSnapshotFailure? How to configure automated snapshot, shall we change the creation frequency?
Configuration:
AWS managed OpenSearch 2.17
r7g.2xlarge.search
3 nodes
EBS: 500 gb
Relevant Logs or Screenshots: