Minute Consumption

Hello there

when i go to the minutes monitoring i always get the error message :

Error - The call for sensor PowerR experienced a timeout. You are probably not in the same local network as the FLM.

my software release is lates and greatest 215

regards

icarus75's picture

Did you enable the local JSON API on the FLM's sensor configuration page?

fl177's picture

mee too, experiencing the
"Error - The call for sensor 5PVrechts experienced a timeout. You are probably not in the same local network as the FLM."

i'm in the same subnet and tried wireless as well as wired connections to this subnet
(netmask 255.255.255.0)
FLM firmware release is v215
JSON API is enabled and have rebooted FLM after setting it.
i tried with several browsers on several operating systems:
safari (OSX 10.6)
internet explorer 9.0 (w7)
firefox 11.0 (OSX 10.6)
firefox (w7)
firefox (w2k)

i only see the timeout error about '5PVrechts' which is my last sensor in the row (on input 5), but
maybe that's because the errors about the other inputs are buried under the error about the last channel.

accoring to the http://flm/cgi-bin/luci/syslog screen, the the inputs (pulses) are coming about every second and the slower curves are drawn nicely

currently no idea how to get the display working with minute resolution
maybe by setting the timeout to a larger value ? where/how ?

icarus75's picture

Hello Johan,

Everything seems to be configured correctly AFAICT. One thing I noticed is that your FLM has a publicly routable IP address. This might mean that the FLM is working within a corporate network. Are your sure the router/gateway is not filtering packets to the FLM's tcp port 8080, which is being used for the local API calls instead of the standard port 80?

The call timeout is set to 4 seconds. This should be ample time for a round trip to the FLM within a LAN environment.

Cheers,
-Bart.

frumper's picture

hi bart,

is the call timeout configurable? i'm getting lots of these timeouts too.

ta,
rik

bazzle's picture

Error - The call for sensor Power consumed experienced a timeout. You are probably not in the same local network as the FLM.

I get this on Internet Exp, and Firefox Win7.

Happens after a couple of minutes.

IE also reports unsecured site on 1st connection.

Bazzle

bazzle's picture

Was getting minute readings yesterday :)
Today, nothing :(
Later today I changed another sensor to pulse (Now called Power Consumed)
Reset unit and reentered settings that worked..
Still no minute readings, Hourly still ok.

Any clues?

Bazzle

bazzle's picture

Hmm... came back in the day time??
Lets see what happens tonight ....

Baz

icarus75's picture

While analog ports will work for minute readings right after an FLM boot, the pulse ports will only start working after they've received their first pulse. This might be the root cause of you not seeing any minute readings at first.

Please note that for minute readings, your browser will be firing one HTTP GET request per sensor every second. If one of those requests fails due to e.g. load on the FLM HTTP server or a network glitch, it will cause a timeout in the browser code. That's when you see the timeout error in the browser window. Refreshing the page will give it a clean start.

frumper's picture

would it be better/possible to just assume a reading of 0 rather than timing out for the pulse ports? some of us have a pulse outputs monitoring solar there is a good chance that the reading will be 0 for 2/3rd's of the day during the winter. refreshing the page does work, but the timeout seems to hang the all the inputs making the minute page less useful than in might be. would it be better/possible to just return the last reading on the pulse for a few seconds to allow for a few retries rather than timing out immediately a packet is lost?

icarus75's picture

The issue will only occur between FLM boot and the very first pulse on the port. From then onwards, the pulse port will return nan's when no pulse has been received for the last minute (e.g. a PV pulse port at night). The nan's will not cause time-outs, but no data will be plotted. The latter is more consistent than just assuming a zero reading.

chrisje59's picture

I see that there other people, more technical than myself have this issue. The question for me therefore is does a clear instruction exist on how to fix this?

bazzle's picture

What I did later was to click the shield symbol on Chrome browser and told it to load unsafe scripts. This worked for local minute readings :)

michi's picture

It seems that the problem with the current minute tab approach is that it is fragile for a lot of people. (There have been numerous posts about this in the past two years.)

I'm wondering whether a more sensible way might not be to have some protocol between the Flukso and the Flukso website to temporarily enable minute tab readings? For example, when requesting the minute tab, that could send a request to the Flukso website that, in turn, instructs the Flukso to start pushing updates to the server at one-second intervals, and then the data would be delivered by the Flukso server. With a time limit, I think this could actually be feasible. (Most people don't spend all day watching their minute tab…)

I know, I'm making things up as I go here. But the current approach seems to cause so many problems that a different approach might be warranted?

Michi.

B-Man's picture

"Bazzle writes......
What I did later was to click the shield symbol on Chrome browser and told it to load unsafe scripts. This worked for local minute readings :)"

do you have to click it each time you load the page or is it fixed every time?
as every time i open it again it doesn't work and have to do the shield symbol trick again

vk2him's picture

Yes you have to click it each time -