Map without stations on it

[BeaumarisWX]

Hi Hans,
In CumulusMX settings : Daily program : At midnight rollover I always have MapsOn.
After last nights test at Midnight the lockfile had success, however the maps.txt that was generated was incomplete.
I have attached all log files including the incomplete maps.txt.

map_midnight_issue_tassie.zip

Kind Regards,

[HansR]

Hi Tony,

In summary: I cannot explain why you have a perfect maps.txt but there are no stations in it. So I blame the activation of two instances (see below), but the proof of the pudding is in the eating. So run again with just mapson and see what happens. And btw, this has no relation with the 3.6.4 beta test, so I'll create a new thread.

One thing I do not understand is why you start two cumulusutils at the same time:
The first is:

Code: Select all

2020-07-04 00:00:03.546 CumulusUtils : Entering Main
2020-07-04 00:00:03.776 CommandLineArgs : starting
2020-07-04 00:00:03.800  CommandLineArgs : handling arg: Top10
2020-07-04 00:00:03.800  CommandLineArgs : handling arg: SysInfo
2020-07-04 00:00:03.800  CommandLineArgs : handling arg: DayRecords
2020-07-04 00:00:03.800  CommandLineArgs : handling arg: Records
2020-07-04 00:00:03.800 CommandLineArgs : End

And the other is:

Code: Select all

2020-07-04 00:00:04.080 CumulusUtils : Entering Main
2020-07-04 00:00:04.104 CommandLineArgs : starting
2020-07-04 00:00:04.104  CommandLineArgs : handling arg: pwsFWI
2020-07-04 00:00:04.104  CommandLineArgs : handling arg: Graphs
2020-07-04 00:00:04.104  CommandLineArgs : handling arg: Yadr
2020-07-04 00:00:04.104  CommandLineArgs : handling arg: MapsOn
2020-07-04 00:00:04.104  CommandLineArgs : handling arg: Windrose
2020-07-04 00:00:04.104  CommandLineArgs : handling arg: NOAA
2020-07-04 00:00:04.104 CommandLineArgs : End

That was intended to be run as one command with all those parameters on the commandline. There is obviously an interaction somewhere but I am not going to analyse that deeply because this is not the way it should be run, even though the logfile system has been taken care of (by the runtime of C#).

What you should avoid anyway is running two instances at the same time (which is what is the case here), that has never been anticipated and apparently leads to unwanted interaction (not understood by me at the moment).

As you run the website, you should run (square brackets means: optional):

Code: Select all

cumulusutils [thrifty] website

But as you don't want that (obviously ), probably because of your modifications in the generated files, you should at least do it in one command:

Code: Select all

cumulusutils [thrifty] all

And then - when this is finished - you run:

Code: Select all

cumulusutils mapson sysinfo

If you want the sysinfo updated - e.g. every hour - you run the last command without the mapson parameter every hour but make sure it won't interfere with the daily command. Because the windrose can take quite some time and you don't use thrifty you should measure once how much time that takes (I see in the logfile for you it is only 3.5 seconds, that will increase over the years!) and account for that in starttime for the next run of CumulusUtils.

So, given your situation, I would advice (e.g.) to do a midnight at 00:01 run (to allow Cumulus to make the rollover to finish if you do not run from within CumulusMX) with the all parameter followed by a maps run at 00:10 and then a sysinfo run every hour at 00:15. That would avoid any unwanted interaction of multiple instances.

[BeaumarisWX]

Hi Hans,
Thanks for the detailed info, greatly appreciated.
I'l modify my setup over the weekend.
Kind Regards,

[BeaumarisWX]

Hi Hans,
I realise your having a break, just thought I'd post this anyway.
Map do not update correctly for me at 00:07 last night, issue with following, : not sure of cause, i'll change my update time again.

Note : Below amended (previous post had incorrect entry).

<link rel="stylesheet" href="https://unpkg.com/leaflet@1.5.1/dist/leaflet.css" integrity = "sha512-xwE/Az9zrjBIphAcBb3F6JVqxf46+CDLwfLMHloNu6KEQCAWi6HcDUbeOfBIptF7tcCzusKFjFw2yuvEpDL9wQ==" crossorigin = "" />
<script src="https://unpkg.com/leaflet@1.5.1/dist/leaflet.js" integrity="sha512-GffPMF3RvMeYyc1LWMHtK8EbPv0iNZ8/oTtHPx9/cc2ILxQ+u905qIwdpULaqDkyBKgOaB57QTMg7ztg8Jm2Og==" crossorigin=""></script>
<style>
#CumulusUtils {height: 600px;}
</style>
<div id="CumulusUtils"></div>
<script>
var CumulusStations = L.map('CumulusUtils').setView([0, 0], 2);
L.tileLayer('https://{s}.tile.openstreetmap.org/{z}/{x}/{y}.png', {attribution: '&copy; <a href="https://www.openstreetmap.org/copyright ... reetMap</a> contributors'}).addTo(CumulusStations);
</script>
<br/><div style ="margin-left:auto; margin-right:auto; text-align:center; font-size: 11px;"><a href="viewtopic.php?f=44&t=17998" target="_blank">CumulusUtils</a> Version 3.9.5 - generated at 12/07/2020 00:07 - &copy; Hans Rottier </div>

kind Regards,

[HansR]

Hi Tony,
If this is the resulting maps.txt (generated at 11/07/2020 17:25) then I do not see a problem. Everybody is on it including yourself as last.
Could you elaborate and send the corresponding logfile?
I'll look at it later when from behind a decent screen, this is from tablet .
Best regards, Hans

Edit: I just saw this nights run indeed did produce an empty map. That is clear. I don't need the resulting generated file, but I would like to see the logfiles of the run itself and those before and after that run. Could you explain how the runs around midnight are organised? Do you still run two commandlines at the same time? You are the only one who has reported this (so far) so I assume it in the way the runs and the results are organised.

[BeaumarisWX]

Hi Hans,
I have amended the earlier post as (I had been storing firewood all day and tired), so must have accidentally sent wrong file.
However I decided the timing had been fine for ages, so elected to leave it set as it was and same happened again this morning.
I wonder if it is something to do with the removal of out of date Stations, as this occurred on both attached file logs.
I forced a manual run just now and it worked OK.

maps.zip

I did a quick check and looks like : https://tillyspaws.com/vp2cutils/index.html also not updated.

<style>
#CumulusUtils {height: 600px;}
</style>
<div id="CumulusUtils"></div>
<script>
var CumulusStations = L.map('CumulusUtils').setView([0, 0], 2);
L.tileLayer('https://{s}.tile.openstreetmap.org/{z}/{x}/{y}.png', {attribution: '&copy; <a href="https://www.openstreetmap.org/copyright ... reetMap</a> contributors'}).addTo(CumulusStations);

Kind Regards,

[HansR]

I wonder if it is something to do with the removal of out of date Stations, as this occurred on both attached file logs.

I am pretty sure you are right, but it is not obvious and I don't see it just like that. So, as it is bedtime here now, I'll have another look later on.
Maybe tomorrow, but I don't have much time then, so it maybe a bit later.

Unfortunately you are the first after the date turns another day so any station which has not given the keepalive timer tick gets kicked off and the first to do that, is where the date difference gets over 7 days. That's always you, down under in Tasmania If you would set it at the end of the day it would be the next in the date line.

The good thing is: every bug has a fix, it just may take some time. As quick as I can.

[BeaumarisWX]

Hi Hans,
Yes I had wondered about first check for the day, sounds about right.
I'll push mine back.
Thanks again, Night Night.
Kind Regards,

[HansR]

I see you added a remark about index.html which I missed last night or did you add it?

Anyway, that apparently is from another command at another time, so I need that specific logfile to see what happened. If I look at your site right now I see an updated index.html so I can't say anything about it from here.

Because it is a separate command and a separate issue, better make a separate message about it or at least indicate EDIT.
Always post the corresponding logfile with it!

I did find the cause for the maps issue, testing.

[billy]

Hi Hans,

Tony is well ahead of me - I have only now noticed my user map has no users, and this thread. An upgrade to 3.9.7 hasn't overcome the problem. If it is of any help, the log file is attached.

2007131552cumulusutils.zip

[HansR]

Hi billy,

Though the result is indeed the same, the cause is really totally different and outside CumulusUtils.

In each step in the process, I download the station sources to make sure it is the latest version (I will optimise that away sometime because now the lockfile protects all individual steps as one single operation but that is basically not so very important right now).

The logfile says:

Code: Select all

2020-07-13 15:53:02.655  DownloadXml: DownloadFile Exception: Error getting response stream (ReadDoneAsync2): ReceiveFailure
2020-07-13 15:53:02.655  GenUtilsMap: XML has not been downloaded, Map has been generated, none or partial Stations on Map!

The issue here is that the last download before creating the map fails. That is not as it should be, but it can happen. Somewhere the internet on its way to the file failed (I think this is the first time I see that). So, I prefer to leave everything as is and I assume a next scheduled run everything will work as planned. If not, no doubt we'll meet again in this thread

[billy]

If not, no doubt we'll meet again in this thread

and we have This error seems to arise no matter what I do - including a reboot of the rpi. From my point of view it is not a big deal - as long as I can still use cutils!

[HansR]

- as long as I can still use cutils!

Yes, as long as a map is generated, no problem. But I dislike is very much because it should not and I like the map very much
I'll send a PM, see how we can handle this.

[BeaumarisWX]

Hi Hans,
Thankyou again, 397 running : https://beaumaris-weather.com/data/index.html
Kind Regards,

[HansR]

Hi Tony,

That's good to hear. We'll watch out for the next station timing out on the keep alive, because that's where the proof of the pudding is
If all stations abide to the minimal cycle frequency, I'll stop updating myself and see if it works (EDIT: I just set my station to non-updating).

Cheers, Hans

NOTE: billy's issue seems to be of a different cause, we're currently working in the sideline. I'll report back here, once we know for sure the cause and the fix.