Weather Display Issue

[EmjayNet]

I'm having a problem with weather display getting information from MH and hoping some of your guys might be able to help.

I've started seeing this error when I start my weather dispaly server in the morning:

"Decoding missed data, raw actual Error: Given start date "20090525201100" is newer than actual date "20090525201012" (all in UTC). Hint: Omit start and end date to get the very latest recording."

I know this means there is a time issue, but Meteohub and the Server WD resides on both get time updates from the same time server, and I've checked both the WD Server Machine and Meteohubs times are out by about 10 seconds.

I don't know what else I can check?

Can anyone offer any suggections please?

[skyewright]

"Decoding missed data, raw actual Error: Given start date "20090525201100" is newer than actual date "20090525201012" (all in UTC). Hint: Omit start and end date to get the very latest recording."

If that is message right at the top of the collection and as a result of it you are not getting the overnight data then try watching the PC's clock before you start WD and aim to start it up just after the change of minute.

I do that as a matter of course, because a long time ago if the change of minute on the PC occurred during initialisation of WD it modified the [Davis Download] time (which stores when data collection was last active) to "now" before collection of logged data started.

I think that was supposed to have been fixed but I stuck with the habit of starting just after the minute just in case...

What you describe could be just that same thing?

[EmjayNet]

Sorry was not to clear.

When WD starts it does the overnight data download just fine, no issues goes through getting on minute data very fast.

I get the error right at the end, and the update window stays open for a minute or two then closes.

[skyewright]

I get the error right at the end, and the update window stays open for a minute or two then closes.

Ah. That's different. It's perfectly normal, i.e. seeing that error message appears to be how WD tells that it has collected all available data from the logger/Meteohub.

[EmjayNet]

Thank you SkyeWright

If that is normal behaviour I can rule that out!

See I don't normally shut my WD Server as it is a internal file, music, & photo server. I use the night time for backups of the various drives, so I had never seen the error message before.

Now just have to work out why WD has stopped getting real time date from MH using logging protocol.

When reading live data WD seems to only get data in batches of 6 (I assume because I have 6 sensors), just stop for a minute or so and clock freezes, than it connects again for another batch of 6, the clock go moves on then freezes again.

[skyewright]

When reading live data WD seems to only get data in batches of 6 (I assume because I have 6 sensors), just stop for a minute or so and clock freezes, than it connects again for another batch of 6, the clock go moves on then freezes again.

Every minute WD asks for "all sensor readings from the last minute".

Every few seconds (around 3-5, I don't recall exactly) it asks for "most recent sensor".

By default, the window that shows data arriving from Meteohub only shows the "all sensor readings from the last minute".

I think you can turn on visibility of the "most recent sensor" data too using

Setup => Advanced/Misc Settings => Log raw data for testing (as rawultimeterdata.txt)

Don't be put off by the rawultimeterdata.txt bit!

You might then need to restart WD? I could be wrong (it's over a year since I turned on whatever it was that activates the extra detail).

However none of that should relate to "clock freezing". Clock freezing in WD is more usually associated with WD having a lot of internal processing to do (file updates and uploads etc.). WD running on my (low power) Fit-PC takes around 5-8 seconds to process the typical "on the minute" data.

Things that take up an especially large amount of WD's time include weather movie processing. Do you do any of that?

Another thought. Last weekend I started trialling a different brand of Anti-Virus software. Till I tweaked its configuration its "on access" file scanner had a really big impact on the length of time taken to handle "on the minute" processing by WD.

[EmjayNet]

Thanks for your help SkyeWright,

I don't run anything other than Windows 2003 file serving, no anti-virus thats in the external firewall.

I'm going to turn on rawultimeterdata.txt as suggested

It's the most recent data thats the problem, my 'Data Received' counter below the clock used to go up by one few second or so.

Now it goes up in batches a 5-6 then freezes, at the same time the clock also freezes, then after about a minute it goes up by 5-6 again, and freezes along with the clock, it just goes on like this.

This mucks up all the graphs and data, the only way I can use WD is the start it up once a day, download missed data then shut down WD again and do that everyday.

[EmjayNet]

OK turned on rawultimeterdata.txt and saved & exit

Restarted, but still the same only gets data every minute, does not get data every few seconds like it use to.

I am really at a loss! So near to throwing the slug out the window :unsure:

[skyewright]

I am really at a loss! So near to throwing the slug out the window :unsure:

Is the system load high?
Does meteohub log show any data problems?
Have you tried Inspect Data => Cleanup Data?

What I'm wondering is if the "most recent sensor" calls are having difficulty. That can be the case with high system load (e.g. when doing recomputation, especially early in the process), and IIRC if there are faulty records in the raw data.

What happens if you use a browser to make calls that are equivalent to the "most recent sensor" calls made by WD? e.g.

http://192.168.1.78/meteolog.cgi?mode=data

Doing that (with you own meteohub's IP) should almost immediately return something like:

20090526202714 wind1 248 7.9 8.1 5.5

[gibsonmb]

THis is a really interesting thread because I still get the problem as quoted below:

"do that as a matter of course, because a long time ago if the change of minute on the PC occurred during initialisation of WD it modified the [Davis Download] time (which stores when data collection was last active) to "now" before collection of logged data started.

I get this about 50% of the time. I then have to go inot the logfiles, check what time/day the data should have been collected from. I then go into the wdisplay.ini file and edit the [davis download] section to the time the download should have started from and save the file then restart WD. Very unsatisfactory. Sp definitely not fixed adn I'm using the latest veriosn of WD (as of about 2 days ago).

[EmjayNet]

Again many thanks for your input skyewright

"Is the system load high?" - System Load: 2.74, 3.16, 2.79
Does meteohub log show any data problems? No last entry was recompution on the 25
Have you tried Inspect Data => Cleanup Data? Yes

"http://192.168.1.78/meteolog.cgi?mode=data"

I get back sensor data, if I keep hitting refresh I get another sensor maybe the same sensor three times in a row then another and another.

I now think the problems is more about WD as you suggested "Clock freezing in WD is more usually associated with WD having a lot of internal processing to do"

Last night I was going through WD control panel screens to see if there is a setting I have chenged by mistake. I open a window from CP (can't remember which one) by mistake it happen to not be full screen so I clicked close and notheing happened. I waited and after about 30 seconds it closed just as the clock moved on and WD got another batch of sensor data.
Tested with more CP windows and it only happens with a couple they freeze when closing and only shut when the clock moves on.

What can make WD do this.

I've looked at the process monitor and WD is not using any CPU 1% every now and then.

[EmjayNet]

Have fixed my issue, data received now counting up every second or so.

Took your advice of Brain on WD forum, went back to a version to WD 10.37N Build 00 as available on WD's download page - older version.


Must be something in later version was causing an issue on my setup.

Will update thread if more info comes to light.

Once again thanks for the help.

[skyewright]

"http://192.168.1.78/meteolog.cgi?mode=data"

I get back sensor data, if I keep hitting refresh I get another sensor maybe the same sensor three times in a row then another and another.

Which is exactly what I'd hope for. So that at least is a :).

I agree that this looks like a "busy at the WD end problem", i.e. not directly related to Meteohub.

It's not a problem I've sufferred from (thankfully!), but I recall seeing that others on the Weather-Watch Forum have.

Things to ceheck out would include:
Do you have lots of file creations and uploads scheduled? If so see if you can spread them out a bit by altering the schedules.

Do you build weather movies?

If WD itself is not unduely busy "on the minute", is something else, e.g. a "on-access" AV scanner? i.e. watch behaviour in Task Manager => Processes sorted by the CPU column.

If those don't help then the Weather-Watch Forum would probably be the best place to search or query in.

Edit: I see you posted while I was writing the above and you have a solution of sorts. :)

[fil]

Hi,

I'm also having this problem, I use meteohub since september last year and this issue has only been happening since last May. http://192.168.0.5/meteolog.cgi?mode=data gives me a blank page after like 1 minute trying to load the page. When WD is switched on, I get 3 or 4 processes in meteohub named wmr928eval using 100% cpu. I also tried different things with WD like using an old version with no luck. :S

[skyewright]

http://192.168.0.5/meteolog.cgi?mode=data gives me a blank page after like 1 minute trying to load the page.

If that happens then it's not a WD issue.

A problem with your Meteohub data is a strong candidate. Have you tried a recomputation? Have you checked the meteohub log for reports of data errors?

David