Hi Hans,

Intermitted link problems (ongoing) here mean I have not been
able to respond exactly when I wanted (not to mention that
nasty subject of paid work ;-).

Analysing the log from your precompiled executable reveals the
following:

TABLE Records IBO FIB
INVENTORY 16101 6.375 16.047
LOGOS 106 21.047 6.063
MUDCOMPANIES 3058 20.485 7.39


Notice the two stand-outs? Those are both tables with large
binary blobs.

Side Note: How did you create those blobs? When I tried
to create an equivalent table using Firebird 2.1 I get
this error from FlameRobin:
Engine Code : 335544569
Engine Message :
Dynamic SQL Error
SQL error code = -204
Data type unknown
Blob sub_types bigger than 1 (text) are for internal use only

I changed the type to 0 and was able to proceed.


So the problem appears to be blobs. So I proceeded to copy the
LOGOS table across to my testing database and ran my two xfer_*
programs against it.

I did not manage to speed up IBO... but I did manage to to slow
down FIBPlus:

= = = xfer_ibo = = =
App Init Time: 16ms
Object Prepare Time: 78ms
Records Transfered: 106
Records Transfer Time: 21328ms
Transaction Commit Time: 594ms
Connection Close Time: 0ms
Application Run Time: 22016ms

= = = xfer_fibplus = = =
App Init Time: 0ms
Object Prepare Time: 188ms
Records Transfered: 102010
Records Transfer Time: 20187ms
Transaction Commit Time: 672ms
Connection Close Time: 0ms
Application Run Time: 21047ms


Isn't that strange? So then I compiled your MaintArcs program
and it shows LOGOS transfer in just 9seconds.

Hans is there any way to validate the logo blobs being
transfered to the server? I keep wondering about that subtype
512 thing and whether FIBPlus and IBO are behaving differently
due to the subtype.

--
Geoff Worboys
Telesis Computing