Sequencing chunk population in dbspace

I am using IDS 7.23 on Solaris 2.5.1. I have defined a dbspace with three chunks and create a table in that dbspace. I then proceed to load the table using a migration utility. How will the engine decide in which chunk to store the data?

Tom Smith wrote: > > I am using IDS 7.23 on Solaris 2.5.1. I have defined a dbspace with > three chunks and create a table in that dbspace. I then proceed to load > the table using a migration utility. How will the engine decide in > which chunk to store the data? It begins using the space in the dbspace from the beginning of the first chunk until that chunk is used up or the required extent does not fit then it proceeds to the next chunk. If multiple tables live in the same dbspace the engine will try to leave unused space between extents of different tables to reduce fragmentation so in that case additional chunks may be started before one chunk is completely full. If you want to spread the load then make the chunks separate dbspaces and use a fragmentation scheme to spread the table's rows across the multiple dbspaces. Art S. Kagel