I am making use of SingleStore Replicate ulitily in snapshot mode to migrate the data from Oracle production to SingleStore cluster.
What would be the impact in Oracle production system for this one time data import ?
What should be the parameters worth considering in the Oracle database so that its not impacted much ( since its production ) ?
It would help if some information is made available about how it works …
I don’t have hands-on experience with this but I will tell you what I’ve heard.
The most common problem I’ve heard about from people who migrated data from Oracle to SingleStore using Replicate is that the source Oracle system can run out of snapshot version space if you are migrating a lot of data and it takes a long time. You can migrate data with Replicate as of a specific point in time and it uses some sort of snapshot versioning on Oracle to do that.
So make sure you have plenty of snapshot version space on the Oracle source. Or if you don’t care about everything being at the same version, you could break the migration in chunks, or do it a different times, not all at one version.
Also, the read queries do put some load on the Oracle source, but that seems to be less of a problem for most people than the version space issue.
Newer versions of Oracle have Automatic Undo Management. Older ones let you set your rollback segment size.
Here’s a related web page:
For snapshot mode we execute multiple select statements on the source Oracle to get the data. Every select query returns a fixed number of rows which is configurable in extractor config file. This config parameter is called fetch-size. If you see default (10,000) fetch-size is impacting performance on source then it can be changed to a lower number to reduce the impact. In my experience at other customers, we have not seen any impact on source Oracle with default fetch-size.