Problem with dashboard graphs

[freddie]

Okay thanks Mark. Worth cleaning it up anyway.

[sfws]

Your log should have an entry "Loaded nnn entries to daily data list" at start-up?

Code: Select all

2020-07-10 08:22:44.567 Loaded 1 entries to daily data list

Similar message for each restart since I updated to what at 3.6.11 was previously latest build. I have not updated to 3.6.12, as I'm wondering about going back to 3.6.10.

Samples from earlier version whose MXdiags I kept, and was working okay:

Code: Select all

2020-07-01 12:40:03.605 Loaded 31 entries to daily data list
2020-06-10 18:39:32.648 Loaded 31 entries to daily data list
2020-06-05 22:11:53.610 Loaded 31 entries to daily data list
2020-04-09 16:32:12.085 Loaded 31 entries to daily data list

Obviously, as I leave MX running when I can, I don't have lots to choose from, but the top one there is version 3.6.10 build where obviously it worked, and the others work back through builds, where it also worked.

The only explanation I have come up with is a system clock adjustment of greater than 1 minute backwards at some point (unlikely I think).

I'm assuming you mean PC clock, not Fine Offset clock.

The PC does update time from some reference site, I am unsure how often, any leap seconds would push it forward not back. As it is a second hand ex-NHS PC, I am unsure how old it might be, but I suppose its internal time-keeping might not be perfect. Since I have stopped and restarted MX without making any difference, I can't imagine any time-keeping issue would stay around.

The Fine Offset relies on a radio signal from Germany and that is very unlikely to skip a minute backwards.

[mcrossley]

Could you post the latest log file in full please?

[sfws]

Having caught up with reading Forum, especially topic viewtopic.php?f=40&t=18289 which has a cross-reference back to this topic, I have updated to 3.6.12 and that has cured the problem. Thank you for fixing it. I see the commit had no typo in build number this time!

[mcrossley]

Having caught up with reading Forum, especially topic viewtopic.php?f=40&t=18289 which has a cross-reference back to this topic, I have updated to 3.6.12 and that has cured the problem. Thank you for fixing it. I see the commit had no typo in build number this time!

That's odd, because that fix was only in the Davis code

[sfws]

That's odd, because that fix was only in the Davis code

I can see that by looking at the changes in the commit, which is why I initially rejected 3.6.12 and considered going backwards to 3.6.10. But since there was suggestion in other topic about Davis that this topic might be related, I gave it a go. My success means some interaction upset Wednesday's build, I'm not skilled enough to see how Davis code might affect Fine Offset.

I felt a bit envious of HansR with his independently generated charts in cumuluutils, and I did even wonder about going back to stable Cumulus 1.

Here is previous diags:

[LAdrain]

Start a new thread please as this is unrelated. I will need to see a log with MX started using the -debug parameter, thanks.

Working fine now with 3088 thanks.

[HansR]

I felt a bit envious of HansR with his independently generated charts in cumuluutils, and I did even wonder about going back to stable Cumulus 1.

That's nice, but all independence is relative. Utils is not the Interface. The dashboard and the charts are basically exactly what you see in the other websites: the code is modified but very much alike and the data is CMX generated. So all not so different at all