Belchertown? It is known Belchertown consumes a lot of CPU time when certain derived values are not in the database, after a 4 => 5 upgrade. Check the group for solutions.
I had maybe something like that a couple weeks ago (without Belchertown), it might have startet after I did an apt upgrade (but no weewx upgrade), I had to restart the RPi4 every day, after 24h I couldn't even ssh into the machine any more. In the logs were messages like "database locked" and those "report thread still running" messages. I didn't care too much because I upgraded my server hardware and installed everything from scratch and don't have any issues any more. I had four weewxd and a lot of other stuff running, I never found out, what caused the issues. Bartosz Francman schrieb am Samstag, 1. Juni 2024 um 15:25:36 UTC+2: > Hi, > For several months I have been postponing the upgrade from version > 4.10.2-1 to version 5.02. To tell you the truth, I didn't even know that it > was a jump from 4.x to 5.x and a few days ago I did an automatic upgrade > (sudo apt update & upgrade). > When asked about saving the configuration file, I left the old version, as > everything worked so far. > My system consists of a RaspberryPi 3B, a Eurochron EFWS 2900 station and > a database (MariaDB 10) located on the NAS so as not to write data to the > memory card. This system has been operating for 6 years (the first entries > in the database are from 2018). So far I haven't had any problems. > Everything worked as it should. > As the weather station only works via WiFi, I had to use the Interceptor > controller to read data from the station. And unfortunately, only port 80 > could be used. It is not possible to configure the station, apart from > turning it on and setting access data to 3 weather services. > Now I will briefly describe the problems with the upgrade. > Port 80 cannot be used by any user other than root, so I had to change the > way weewx runs from the weewx user to the root user. Data began to appear > and the script did not end with the lack of access to port 80. I thought > that it was not so bad. > However, I saw that the processes related to weewx were using 100% of the > load of two raspberry processors. I found information on this group that it > may be a problem with the lack of appropriate columns in the archive table ( > https://groups.google.com/g/weewx-user/c/6rl2FIbqVp4/m/S0Ek9ZVaBwAJ) I > did everything according to this post, plus from > https://weewx.com/docs/5.0/utilities/weectl-database/#update-a-database > update the database and check it. All this took about 2 hours due to the > large amount of data I have in this database. By the way, apart from the > high CPU load, I noticed that every half a minute there is a very large > data read from the database. And it's related to weewx. I thought that > rebuilding the database would reduce the CPU load and eliminate the > frequent polling, but unfortunately it didn't help. I don't have any very > advanced visualizations, just a basic diagram. > I left it because there is nothing else to do with this raspberry and two > out of four processors can be fully loaded. But unfortunately, about two > hours after starting the system, I received information from one weather > service that I was not sending data to them. I tried to access this > raspberry via ssh, but it was impossible. Connection timed out. Although > the raspberry was running. > Only today I sat down to work on it and managed to connect to the > raspberry. The last log entries after issuing the sudo systemctl status > weewx command are provided below: > > pi@raspberry-pi:~ $ sudo systemctl status weewx > × weewx.service - WeeWX > Loaded: loaded (/lib/systemd/system/weewx.service; enabled; preset: > enabled) > Active: failed (Result: signal) since Sat 2024-06-01 10:36:12 CEST; > 1min 2s ago > Duration: 19h 10min 2.367s > Docs: https://weewx.com/docs > Process: 720 ExecStart=weewxd /etc/weewx/weewx.conf (code=killed, > signal=KILL) > Main PID: 720 (code=killed, signal=KILL) > CPU: 1h 59min 21.575s > > maj 31 17:51:09 raspberry-pi weewxd[720]: INFO weewx.restx: PWSWeather: > Published record 2024-05-31 17:48:00 CEST (1717170480) > maj 31 17:51:22 raspberry-pi weewxd[720]: INFO weewx.restx: > Wunderground-RF: Published record 2024-05-31 17:48:47 CEST (1717170527) > maj 31 17:51:28 raspberry-pi weewxd[720]: ERROR weewx.restx: WOW: Failed > to publish record 2024-05-31 17:48:00 CEST (1717170480): Failed upload > after 3 tries > maj 31 17:51:28 raspberry-pi weewxd[720]: INFO weewx.restx: > Wunderground-RF: Published record 2024-05-31 17:49:19 CEST (1717170559) > maj 31 17:51:53 raspberry-pi weewxd[720]: INFO weewx.restx: > Wunderground-RF: Published record 2024-05-31 17:49:51 CEST (1717170591) > maj 31 17:52:13 raspberry-pi weewxd[720]: INFO weewx.restx: > Wunderground-RF: Published record 2024-05-31 17:50:23 CEST (1717170623) > maj 31 17:52:56 raspberry-pi weewxd[720]: INFO weewx.manager: Added record > 2024-05-31 17:50:00 CEST (1717170600) to database 'weewx_metric' > cze 01 10:36:12 raspberry-pi systemd[1]: weewx.service: Main process > exited, code=killed, status=9/KILL > cze 01 10:36:12 raspberry-pi systemd[1]: weewx.service: Failed with result > 'signal'. > cze 01 10:36:12 raspberry-pi systemd[1]: weewx.service: Consumed 1h 59min > 21.575s CPU time. > > Yesterday at 17:52:56 there was the last entry to the database and since > then "nothing" has happened. Only today at 10:36:12 weewx stopped. When I > logged in. Yesterday I tried to log in around 18:00 and the system was > frozen. > While checking the logs, I found this information regarding weewx version > 5.0.2: > > WARNING weewx.engine: Previous report thread has been running > 601.9695854187012 seconds. Launching report thread anyway. > > INFO weewx.imagegenerator: Generated 13 images for report SeasonsReport in > 458.25 seconds > > INFO weewx.engine: Launch of report thread aborted: existing report thread > still running > > INFO weewx.cheetahgenerator: Generated 8 files for report SeasonsReport in > 252.10 seconds > > Generating simple images shouldn't take more than 30 seconds of time, > right? And not, as here, over 7 minutes in one case and 4 in the other. > In version 4.10.2, generating the same images takes: > > INFO weewx.cheetahgenerator: Generated 8 files for report SeasonsReport in > 12.51 seconds > > I came to the conclusion that this was no longer enough for my nerves and > I went back to version 4.10.2-1. Now the system works ok, nothing loads the > processor, data is sent to weather services on an ongoing basis. The local > page is generated every two minutes. > In version 5.02. the page was refreshed at various intervals, when I > entered, I had data from 20 minutes ago. > > Below I am posting debug information from my settings for 5.02 and a few > drawings illustrating the problems. Maybe someone will be able to find out > what is wrong in my configuration... Or maybe the problem is somewhere > else... > > BR > Bartosz > > Debug file: https://gofile.me/75BgA/0SlVLSvVO > > CPU load 5.02 > [image: weewx.png] > > CPU load 4.10.2 > [image: weewx_4.10.2.png] > > db load 5.0.2 > [image: db.png] > > db load 4.10.2 > [image: db_4.10.2.png] > -- You received this message because you are subscribed to the Google Groups "weewx-user" group. To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/weewx-user/456480dc-23ab-4d93-ae82-500a277a9d2en%40googlegroups.com.
