[ts-gen] Known IDEALPRO Outages

[Ken Feng]

I thought I would start a new thread listing known Forex Outages to
help out other users who are trying to keep TWS/shim going as long as
practicable.  These outages may also apply to other markets, but I
suspect if IB was smart about this, they tried to pick small windows
where no other exchanges are open.

This is what I know so far, after playing with TWS/shim for a couple
of weeks.  I will update this thread as I find out more over time.

ShimText is your friend - learn to read it and help you diagnose problems.

The goal is to start TWS/shim at the beginning of the week and have it
run continuously for the entire week, without intervention.  TWS/data
quality is a separate, but no less important issue.

I hope others find this information useful.

- Ken

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

This is using servers: usfarm, cashfarm, ushmds2a

UTC+9 is Tokyo time, what I use...

IDEALPRO "EXCHANGE" HOURS
# --------------------------------------------------------
# WEEKLY
# ON  AT SUN 21:15:00 UTC OR MON 06:15:00 UTC+9
# OFF AT FRI 21:00:00 UTC OR SAT 06:00:00 UTC+9
# IB IDEALPRO 17:15 - 17:00 (ET) Sunday-Friday
# --------------------------------------------------------
# DAILY
# FIRST TICK 21:15:00 UTC OR 06:15:00 UTC+9
# LAST TICK  20:59:55 UTC OR 05:59:55 UTC+9
# --------------------------------------------------------

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

OUTAGE 1
20:59:55UTC or 05:59:55Tokyo is the last tick of the day
21:15:00UTC or 06:15:00Tokyo is the first tick of the day
OUTAGE 2
03:55:00UTC or 12:55:00Tokyo (approx) TWS ok, but bars are no longer sent
04:05:00UTC or 13:05:00Tokyo (approx) TWS ok, can receive bars again

Workarounds:
OUTAGE 1 - not needed, as long as TWS/shim is up, shim will wait until
TWS starts sending bars again
OUTAGE 2 - a bit tricky - upon the first "Market data farm connection
is OK:cashfarm" message, you want to cancel your bars and re-request.
For about 10 minutes you may get various broken and connect messages.
Ignore them.  After a ten minute period, everything should be up and
running again.