[ts-gen] order execution and connectivity problem

[R P Herrold]

On Tue, 9 Oct 2007, Nils Gebhardt wrote:

> market opening was 30 minutes before the order command was logged.

Then this is a clear problem which IB has if on a 'live' 
account.  I do not know for certain other than by observation 
as to 'paper' accounts [which are not the 'edemo' account], 
but we get seemingly identical fills here from the paper and 
'real' accounts as to 'MKT' orders.

>> You should be able to use this email and the command.pdf
>> reference as a template to see into the right time frame of
>> History.  Please let the list know the result of your
>> investigation.

> I can count 588 trade events in the time window in question.

well, that settles that -- something is really wrong.  If we 
can get a reproducer within the shim, we'll fix it of course; 
I suspect that it is upstream, as the 'bin/unsafe' test script 
seems to cover the same features as your example quoted 
previously.

> Actually I had the same problem in the same set up the first 
> time some days ago already, but that time I was rather 
> slightly confused than worried. I think the time gap was 
> even worse (>100 min).  Unfortunatly I do not know how to 
> reproduce the scenario, usually it runs fine.

IB TAC may be interested in your logfiles, as well.

> I just picked from commands.pdf the commands 'verb' and 
> 'open', that might help to get more output in case the 
> phenomena reoccurs and to clarify weather tws acknowlegded 
> the output of trading shim.

I have not tested the 'open' command for a long time, but know 
of nothing other than bitrot which would have affected it.  ;)

'verb' works fine and is tested continuously daily in my 
regression scripts.

Thank you for the prompt reply.

-- Russ Herrold