Keith Fang (fang@netcom.com) wrote:
: >The chunks are known as primary chunk for the first chunk
: >in a dbspace and secondary chunk for any additional chunks added.
: >(This is not an exact quote, to avoid possible copyright infringements.)
: >
: >Bunk. Bull. Incorrect.
: >
: >Do a tbstat -d. Notice the flags. Do you ever see a flag of "SO-"?
: >No, you will see either "PO-" or "MO-". The "P" means "PRIMARY", and
: >the "M" means "MIRROR". Any chunk, I mean ANY CHUNK that is not a mirror
: >chunk is a primary chunk, regardless of order of creation. Informix has
: >no such concept as a secondary chunk.
: Try dropping chunks in a multi-chunk dbspace. You'll get the error:
: Chunk not dropped.
: ISAM error: Can not drop first chunk
: when you try to remove the first chunk.
Of course you'll get that error. The first chunk of a dbspace contains
the tblspace tblspace for that dbspace, so you can't drop it. I never
claimed that their are no differences in the chunks, just that there
is no such thing as a secondary chunk.
I quote from the class guide for the "Informix Internal Architecture
and Advanced System Administration" class....
Myth Busting: The term "primary chunk" is often misused. It means
"non-mirror chunk". It does NOT mean "root chunk" or "first chunk in
DBspace".