Pi 4b extremly slow running openplotter 3

[baltika_no_9]

OK if it's slow immediately after reboot then that's good, we know it's probably not a memory leak.
The white screen is odd, how about turning off the display while you're not there?

Her's a link for adjusting the swap file size if you need to do it https://pimylifeup.com/raspberry-pi-swap-file/

[amckeown]

Tried turning off the influxes, and it seems to speed up a bit. For some reason, i cannot find the config.sys file to change the swapfile size. Searched several times, and the find tool doesn't find it. Left it running the other day, to see if a long time works. heading down later today.

When I looked yesterday there was about 4 mb disk space left. Have a new, larger samsung plus card I can copy the card too.
It is possible I set the influxes up so that excess data is written. Have a 7in sunfounder touch screen. Have not been able to get in to have a look at the setup for data.

I suppose the sd card might have got damaged. When the Pi "locks up" the only way i can get it to reboot is pull the power. ( its in a pican M hat with power from the network).
wondering if it would help if I upgraded to the new version of open plotter? if so, is there a way of copying the influx setup to the new version?

FYI, the network carries: gps position, water depth, temp, speed; fluxgate compass direction, masthead air temp, pressure, RH, wind speed wind direction:, and internal cabin air temp, RH, air pressure via ic2 input.

[baltika_no_9]

You are looking for the wrong file, it isn't config.sys, it's /etc/dphys-swapfile and it has to be there, it's part of the standard build. However, if all the space you have left on the card really is 4mB (in post #1 you suggested that there was 20GB free, now you have only one five-thousandth of that) then increasing the swapfile is not going to help anyway.

If you really are storing a lot of data then even a bigger card will be filled up too in time. Did you set a retention policy on your Influxbd data? Maybe look at the documentation for this depending if you're using version 1.x or 2.x.

How about installing baobab, that will show you what's using up your SD card storage. Changing to a different version of OP is unlikely to help if you're trying to store too much data, much of it redundant (position, speed, direction etc aren't going to change if the boat is tied up.)

[amckeown]

Thanks for the reply.

Sorry about the delay. Too hot to be in the boat cabin for my liking. last about 10 min.  Thunderstorms again this morning so I'm getting caught up on emails.

3 gb free on the chip. its a samsung evo plus, 32 gb. Aside from being slow, the system check starts when clicks and disappears. Wanted to check the influx variables setup, and cannot get into it. my best guess is I made an error in the set up for retention time.

ran the card diagnostics program, see results. 
  rpdiags.txt (Size: 1.07 KB / Downloads: 18)

notice a canjs error reached past the end of a stream, not sure what that means     .
       

on one try to get into the local host got not found.

these were with chromium, same in firefox.

tried to get into influx to look at data. I get the screen, but cannot do anything with it.

going to try the suggestions.  I want to figure this out.


could not find the swap file either. search was left on overnight.

wondering if something has gone wrong with the chip? when I get a chance going to pull the box the pi is in apart, and put in a backup chip.  and see if i can get into the data setup. I can see if I can change the retention times for some things, want to keep the weather data .

Do have a samsung ssd I could put in and run openplotter from it. That was on my list to do on the pi this summer.

[baltika_no_9]

"on one try to get into the local host got not found."

The fact that you can't get into localhost:3001 suggests that grafana is not running

"could not find the swap file either. search was left on overnight."

You're not looking for the swap file so searching for it is of no use (it's at /var/swap), you're looking for the config file /etc/dphys-swapfile. It's part of the standard build. The following command should find it directly.

Code:

pi@folly:~ $ ls /etc | grep dphys
dphys-swapfile

The fact that there are failures shown in your SD card diagnostics should be no surprise as that test is intended for new cards. See the caveat:

 "The pass / fail criteria in this test assume a new (or at least freshly formatted) card; don’t be alarmed if the write speed test fails when run on the SD card you’ve been using for the last six months"

The write speeds shown though are woefuly low, As mentioned earlier a faster card might help.

[baltika_no_9]

You could check how big your data buckets are (assuming you are using influxdb 2.x). If you really do have a lot of data consider offloading it and handling it offline in your own time. Then clear the bucket(s), that way you won't lose any data but you'll free up space on the disk. Do you really need to keep it all on the Pi?

[amckeown]

Finally, a cool day and I spent time with it. Yes i'm using influx 2.x. I'd move the data off if I knew where it was stored. One reason i bought the t7 ssd. I'm used to windows where one can route data anywhere you like within reason. Spent a lot of time with the manual waiting for a cooler day.
Been through the bucket for influx db, and was only storing environmental data, inside outside temps, rh, pressure, water temp, wind speed direction, , no positions etc. Unchecked a couple I really don't need.
Rebooted, and inadvertently started chromium. Talk about slow. Tried to work with influx db but unworkable. Only thing that works with chromium running is the instrument panel.
Made sure firefox esr was set as the default browser. File manager worked ok today with firefox running. Check system as well.

[baltika_no_9]

Data is stored in /var/lib/influxbd/engine/data and you can see how much space the data is occupying with the command:

Code:

sudo du --max-depth=1 -h /var/lib/influxdb/engine/data

Here's mine, I have only one bucket called test (apart from the system generated ones)

   

you'll see it uses bucket ID and not bucket name. To cross check, go into the influxdb GUI

   

There your see bucket name and ID, also the retention period. In settings you can change the retention period and delete data but if you want to keep your historical data then don't do that just now or it will be lost.

Now if you go into Influxdb Data Explorer, select the measurements you are interested in, select the format (raw data or graph), period and then submit.

To download this data to a csv file then just press the relevant button shown on the screenshot. BUT I suspect if you try to download all measurements for all time it will blow up (you'd be trying to store another copy of the data to your Downloads folder on an already nearly full SD card). Either download the measurements separately and for a defined period, say a month each and then move the resulting csv file to separate USB memory then delete from the downloads folder. OR clone your system to a bigger SD card/SSD then download the lot. And do this periodically so you have all your data saved offline and it doesn't clog up you card. You can upload any of these downloaded csv files back into Influxdb should you need to or just view the historical data on another system (Windows desktop etc). Another option and one I like to use as I prefer CLI to GUI is to download the data directly to a separate USB memory device so you're not writing to the SD card. To do that I think you'll need to set up a profile configuration which I suspect you haven't. It is quite easy.