Stay up to date with the latest from the Knowledge Center. See all new Knowledge Center articles published in the last month, and re:Post’s top contributors.
Variation in Export task start time
Dear community,
I have question regarding Exporting data from Aurora snapshot to S3.
Until recently, everything worked fine - Lambda would trigger creation of Aurora PostgreSQL database snapshot, and then when creation of snapshot is finished it would trigger exporting task to S3 of snapshot data.
Recently, it takes much longer time for export task to actually start, even 5 hours after snapshot created event trigger. Its stuck in STARTING status for long time and after that partial export finishes in approx 10 minutes. Why is this happening?
- Newest
- Most votes
- Most comments
The delay you're experiencing with your Aurora PostgreSQL snapshot export tasks being stuck in the "STARTING" status for extended periods (up to 5 hours) before quickly completing is not uncommon and can be attributed to several factors:
-
Snapshot Restoration Process: When exporting data from a DB cluster snapshot, AWS first needs to restore the snapshot before extracting the data to Amazon S3. This restoration phase is what appears as "STARTING" in the status. The export task first restores and scales the entire database before beginning the actual data extraction to S3.
-
Network Traffic and AWS Region Bandwidth: Restore times are significantly impacted by the amount of network traffic an AWS Region receives relative to its available bandwidth. When there's a sudden increase in traffic in your region, you might experience longer than expected completion times for the starting phase.
-
AWS Service Prioritization: Since snapshot exports are asynchronous processes, AWS may prioritize different operations at different times, which could explain the recent change in behavior you're observing.
-
No Performance Impact Design: The export process is designed to run in the background without impacting active DB clusters, which sometimes means it gets lower priority than other operations.
For faster exports, you might consider using live DB cluster export to S3 as an alternative. This approach has shorter start times than DB snapshot export because it doesn't need to restore a snapshot first.
If this issue continues to be problematic, you can monitor the export tasks more closely using the AWS CLI's describe-export-tasks command to get more detailed information about the status and progress of your export tasks.
Sources
Creating snapshot export tasks - Amazon Aurora
Exporting DB cluster snapshot data to Amazon S3 - Amazon Aurora
The delay in export task starting (STARTING status) is normal behavior as the process involves restoring and scaling the entire database before data extraction to S3 begins. Once the initial setup is complete, the actual export (IN PROGRESS status) proceeds quickly.
Key points:
- Export time depends on database size and type
- Process involves full database restore before export
- Status flow: STARTING > IN PROGRESS > COMPLETE
For troubleshooting:
- Check CloudWatch logs for StartExportTask API patterns
- Verify if multiple concurrent exports are running
- Contact AWS Support if delays persist
Source: AWS Documentation - Export Snapshot to S3
For detailed troubleshooting: AWS Knowledge Center - Aurora snapshot export issues
Relevant content
- asked 4 years ago
- asked a year ago
