hypertrend trace sloped or stepped?

[Ryan.S]

Yes, I notice that there is no line when it is zero. That's why I wonder if the communication is good after it is zero.

I mean that if lookout doesn't get more data after zero, it should not draw the line. For example, the last zero point is at time 10:00. If lookout doesn't get more data for 10 minutes, there will be no line from 10:00 to 10:10. If lookout gets data 0 again at 10:11, the "ghost point" will be updated, and the hypertrend will draw a line from 10:00 to 10:11.

But if lookout gets data 5 at 10:11 directly, the data 5 will be logged to database. Then there are only two points in database, 0 and 5. The hypertrend will just draw a line to connect 0 and 5, because it doesn't know what happens from 10:00 to 10:11.

That's why I want to have a look at the communication. If the communication is good after it is zero, and lookout gets updating data, the hypertrend should draw a line. But if there is no incoming data, the hypertrend should stop because it doesn't get updated data from database and the "ghost point" is not updated either.

[Ryan.S]

Note that the data in database and on hypertrend represents exactly what lookout driver object gets from device. So, the MOSCAD problem is easy to explain. During the delay of 30 minutes as you said, I assume that there was no data getting to lookout, so lookout didn't log anything and the hypertrend stopped updating.

This is just my guess of what happened to your opclient object. Maybe I'm right, maybe not.

[dc@aci]

Ryan,

The value of 0 flow is continuously updated.  I determined this by setting the value to zero in Lookout and then watching it update again from the OPC driver.  The data stored to citadel is ONLY stored to citadel when the value changes from zero.  This is the problem.  The flow OPC issue is perplexing since during the entire time the value is zero it does not log anything to Citadel.  I can understand why since the purpose is to report by exception.  This is fine but it appears that the algorithm for trending is adversely affected by the assumption that the last value of zero occurred when the last value of zero was logged to Citadel.  In realitiy, it occurs just before the value has changed and therefore, the trend is inaccurate.  I hope I am explaining this clearly enough.

I will get the diagnostics to you as soon as possible so you can review.