Blob threadpool

[Mark Addleman]

I'm writing an application which may write tens of thousands of blobs simultaneously.  Each blob will be around 1k. 

In initial testing, it appears that H2 writes each blob using a separate thread.  I doubt that approach will scale in my use case.  While I could adjust the design of my app to accommodate this behavior, I'd prefer some direct control over it such as providing my own thread pool or something else.

Does BINARY / VARBINARY have this same behavior?  Maybe that would map better to my use case.

[Noel Grandin]

With a BLOB that small, you'd be better off using a larger PAGE_SIZE, and a VARBINARY column.


Mark Addleman wrote:

--
You received this message because you are subscribed to the Google Groups "H2 Database" group.
To view this discussion on the web visit https://groups.google.com/d/msg/h2-database/-/2oJVTsrQglYJ.
To post to this group, send email to h2-da...@googlegroups.com.
To unsubscribe from this group, send email to h2-database...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/h2-database?hl=en.

[Thomas Mueller]

Hi,

> separate thread

The separate thread is only created when calling
Connection.createBlob(), and then Blob.setBinaryStream().

For H2, PreparedStatement.setBytes is a lot more efficient. Or
PreparedStatement.setBinaryStream (if you don't know the length, use
-1, or the method without the length when using Java 6).

> With a BLOB that small, you'd be better off using a larger PAGE_SIZE, and a
> VARBINARY column.

I agree. 1 KB isn't large, and the BLOB overhead is most likely larger
than the benefit.

Regards,
Thomas

[Mark Addleman]

That's very useful to know.  Thanks