First published on MSDN on Jul 25, 2017
The Azure SQL DataSync workflow is carried by different tasks at different stages. The first task is to provision destination databases (...
Updated Mar 14, 2019
Version 2.0
Blog Post
You mention that reprovisioning is resource intensive. Against which database is it intensive against?
Typically we have 3 databases in our sync process. There is the AzureSQL database, the HUB database (resides in AzureSQL, but different to the data source) and also an OnPremise database. Which of these DBs would need additional resource to improve reprovisioning performance? There is also a VM which holds the Microsoft Data Sync 2.0 Application. This box is very low spec and seems to do very little throughout a sync. Is this normal?
I have the same question regarding the Initial Synchronization.
When performing rebuilds of this, I find that this is quite slow and am wondering where the bottleneck might be:
1) AzureSQL database
2) Hub Database
3) VM that holds Microsoft Data Sync 2.0 Application
4) Destination location (On Premise SQL Database)
I understand that improving HUB performance will make the GUI more performant, but once we press the SYNC button, it seems to just chug along at its own pace, with nothing looking particularly taxed, yet it still takes 14-16 hours to sync. (During which time we have zero way of monitoring progress).
Is there any particular area I could look at to improve performance on to improve the sync speed?
Thanks