Cumulus Support

n9mfk wrote:when i did that it crashed and told me to post my mxdiag

Some kind of problem with your serial port (or the Silabs drivers, if it's actually a USB weatherlink). You may need to reboot.

2015-11-13 14:32:00.370 Serial device = COM1
2015-11-13 14:32:00.371 The port 'COM1' does not exist.

jpsc wrote:For what it's worth I've seen this blank control panel, with just the time updating. I went through to the settings and changed some things before stopping, rebooting, and restarting. Sorry, no diagnostics.

I have had this happen more than one time, and on more than one browser.

I corrected it by clearing the cookies on the site. Works for me every time.

Hi Steve,
after reboot it started here are the files

Yes, you're back to the same problem as before. Something apparently happens when it tries to write the yesterday.ini file for the second time which causes the logger reading thread to hang. I don't know what it is and I can't reproduce it myself. It seems no-one else has this problem - something on your system is apparently causing the writing of the file to hang.

Until we can work out what the cause is, you'll just have restart MX when it gets to this point, it should then carry on as normal.

ok let me known if there more test i can do to help or if you need my files to test
Beau

After the next build, I think. I'll add some more logging, and look at changing to the later version of the ini-file library.

Hi Steve i tried to upload a diag but it was too big
https://www.dropbox.com/s/orcq21mhqs0m2 ... 9.zip?dl=0
did it load ok

That's eliminated the writing of the yesterday.ini file as the cause. I think that what might be happening is that the station has gone to sleep while Cumulus is doing the end of day rollover and creating the NOAA reports (seems to take about 12 seconds on your PC), or perhaps rather than going to sleep, it has forgotten that it was supposed to be sending archive data.

I think I'll have to add some more logging to try to get closer to what the problem is. I'll also add some code to handle the console stopping sending archive data, which might solve the problem anyway.

I actually hit this problem myself just now. We had a power cut overnight, and the UPS eventually ran out of power, so my Raspberry Pi shut down. When I started it up again, it had about 90 archive entries to download, and there was an end of day reset in the middle (mine is at 0900). It stopped processing archive entries at that point. I realised that I wasn't running my new code yet, so I copied that over and 'rewound' to force it to repeat, and the problem was indeed that the logger had got bored waiting for MX to do the end of day rollover, and stopped sending archive data. My new code detected that, and the second round of downloading archive entries then ran to completion.

So I'm hopeful that the next build will fix your problem, finally.