Cumulus Support

Hi there!

Within the past few days I found that when I access ...8998/index.html on my local subnet, the page loads fine, but does not pick up any data. I restarted MX to no avail, then rebooted - also to no avail. I've looked at the logs I know about and could not spot anything concerning. (Seems I can access all the settings and such).

I can't think of anything that has changed recently. The Pi 2 has been up for~ 260 days, and Cumulus was restarted a couple weeks ago when it stopped talking to Wunderground. I have some monitoring scripts or log from them I may have touched, but there's been no reason to tinker much of late - so I'm stumped. (I know ... people always say "I didn't do nothing." I've ssh'ed into the pi twice over the past week, most likely to check the Diags log.)

I'm running old stuff for years at this point. It's been working fine except for Wunderground burps. I am running Cumulus MX v.3.0.0 build 3043 under Jesse. I've written a fair amount of Python code - some to monitor CMX and the Pi - but I've messed with the web side very little, and certainly not recently. e.g. I tweaked the footer of interface/index.html in April.

Any suggestions on how to debug this? - - - Of is this a head-slapping "I know exactly what this is?"

Thanks!!
Bob

Update!?

Enjoy,
Paul

I'd have a look at your browser dev tools - particularly the network tab.

OK ... it is a head slapper, dang it. I reloaded the page and happened to see something about cache out of the corner of my eye.

I confess I resist flushing cache because it is a bit of a nuisance for a little while, so I tried a Chrome Incognito window and voila.

Thanks for the replies though!

... I've never dug into it, but it still suprises me how often flushing cache fixes things. There were no server (or client?) -side changes, and this server has literally been working for years without ever seeing this issue. That suggests to me that something external to CMX was likely the cause. I've QA'ed a lot of software before retiring and browser-based pages seem to me to have some level of (unexpected) fragility. That's just the way it is I suppose.

Stay Safe Out There-
Bob

PaulMy wrote: ↑Sat 20 Jun 2020 4:44 pm Update!?

Yeah, I know. I hear you. I know you guys (Paul and Mark) contribute a lot to CMX. I've updated quite a few times before this version, and from a previous version of Cumulus, keeping old data. I believe I've also accumulated a few outdated files over time. I've had in mind to take an inventory for all my CMX files, load a new version in /tmp or something and basically diff them to determine which files are no longer needed.

Actually the better answer would probably be to run the new CMX for a few minutes to accumulate a little data to make it a better comparision. And of course read through all the "release notes" to understand what else I need to account for. I do recall the SteelSeries ... err HighStock stuff changed at some point. Glad that Mark (et al?) is keeping CMX going.


Truthfully, because I have level of comfort with Python, and if I want to do a careful update and incur some effort and possible breakage, investigating WeeWX has been in the back of my mind and may make sense. I ran Cumulus under Windows for a long time before CMX and I understand why mono is required, but the few times I tried to move off v 3.2.8 I couldn't get CMX to work. Having done support too during my career, I often try to run the lastest stable version of most things, but not here.... *sigh*

Bob . . . . sometimes mr. if it ain't broke don't fix it. (This was a close one)

Idem for me, more synchronization from ....: 8998 / html index, while cumulusMX is well started as administrator finds the central unit etc.

Last Cumulus update RED
it worked until the 15th of this month without problem despite the updates made systematically

Help me ?