Export Azure DataBase

Question

Export Azure DataBase

Jose S 20

I have a database that is approximately 180 GB in size. It had never been exported to a Storage Account before. Last night, a manual export was initiated at 05:30 UTC, but the export process did not actually start until 13:00 UTC. My question is: why was there no activity during those nearly eight hours?

0 comments

Answer accepted by question author

Jose Benjamin Solis Nolasco 12,691 Volunteer Moderator

Welcome to Microsoft Q&A,

@Jose S I hope you are doing well,

For a database of approximately 180 GB, the delay between submitting the export can be related to the Azure SQL Import/Export service processing the request before the actual export work starts.

More importantly, Microsoft currently recommends SqlPackage for databases larger than 150 GB https://erikej.github.io/efcore/sqldb/sqlpackage/2022/08/05/azuresql-sqlpackage.html. Portal/PowerShell export requests are processed on Microsoft-managed machines with 450 GB of local disk, and the temporary DacFx files can require up to 3× the database size. For a 180-GB database, that can exceed the available local disk capacity and cause delays or failures.

For future exports of this database, I would use SqlPackage from a VM or other machine with sufficient local SSD storage, preferably in the same Azure region as the database.

Microsoft also notes that export performance depends on workload and database structure. Temporarily increasing the database compute tier, minimizing concurrent workload, and ensuring large tables have clustered indexes can improve export performance.

References:

https://learn.microsoft.com/en-us/azure/azure-sql/database/database-export

https://learn.microsoft.com/en-us/azure/azure-sql/database/database-import-export-azure-services-off

https://learn.microsoft.com/en-us/sql/tools/sqlpackage/sqlpackage

If this answer helped clarify the platform capabilities and save you troubleshooting time, please consider marking it as Accepted. This helps others in the community find similar solutions.

0 comments

1 additional answer

Your answer

Answer 1

That gap is typically queue time in the Azure SQL Database Import/Export service, not active export time.

The portal-based export uses a regional REST-based Import/Export service that has a limited number of compute VMs available per region. When many import/export requests are submitted in the same region, requests can wait in the queue before resources are assigned. During that period, little or no visible activity occurs. The documented delay can range from seconds to many hours.

For a 180-GB database, this is more likely to be noticeable because large BACPAC exports are long-running logical exports, and the service can also be constrained by temporary disk space on the machines that process portal export requests. Microsoft notes that these machines have 450 GB of local disk space and that import/export can require up to three times the database size in disk space. For large databases, the recommended workaround is to use SqlPackage on a machine with enough local disk space, and databases larger than 150 GB should use SqlPackage instead of the Azure portal export path.

If more predictable start times and throughput are needed:

Use SqlPackage instead of the Azure portal export.
Run it from a VM in the same Azure region as the database.
Temporarily scale up the database compute during export.
Reduce read/write activity during export, or export from a transactionally consistent copy.
Ensure large tables have clustered indexes, because large heap tables can slow or fail export.

Also note that if an export operation runs too long, it can be canceled by the service.