[evla-sw-discuss] C125660

[Ken Sowinski]

On Mon, 5 Oct 2009, Pat Van Buskirk wrote:

> How should we handle idcaf errors during Widar tests?

Ignore them.  In my limited experience idcaf and the old
correlator sometimes get confused about what they are doing
when Widar is in use and we change the number of antennas
in the non-Widar subarray.  As long as the primary puspose
of Widar testing is to test Widar, then we ignore these
problems since idcaf/telcal/d10 is not a part of the Widar
data stream.  In this case, as it was explained to me, the
alert from idcaf did not appear until the dynamic observation
had finished and the non-Widar antennas and old correlator
were idle.

Suggested action:
1.  If idcaf alerts appear at the end of a non-Widar script
check to see if the number of antennas in the subarray has
changed.  If so, try again with the same number of antennas.

2.  There is never any need to kill/restart idcaf or executor
if there are no other symptoms.

3.  If the above is not enough, wait till the Widar tests are
done and the system will again work when all the antennas
are using the old correlator.

Ken