as starting with a clean installation on my Rpi5, there was no indication for debugging. In parallel I monitored the activitation with s-tui, especially for cpu load/temperature and fan speed. So I started with simulator modus, then installed gw1000, all of it without any increase in cpu activity. Next day, I did the adoption of alltimeSeasons, as this needs to be installed completely manually. By this way, I had to correct two gw1000 specific entries, concerning the transmitter battery and the sun irradiation, both of them were empty. I transferred both entries from the old Rpi4 conf file to the new one, again without any change un cpu usage. Finally I installed windy upload. Still no problem. So, the only two things, I did not install, is forecast and cpu-temp. The latter, however, raised some cosmetic trouble as the data were collected in extraTemp1, which is now used by a meanwhile installed barn sensor for temp and humidity. I should try to delete the old data for cpu-temp, because it looks crasy, when temp was about 50 °C and now dropped to 10 °C. As I monitor the two Raspis now in Home Assistant, I will not install this feature in weewx again. Concerning forecast, I am still thinking about.. Still, I doubt, that this is the cause for the high cpu usage at the Rpi4.
Before I decide this, I could use the Rpi4 installation now for testings. Maybe it is a good idea to remove one part by the next of all the additions, and look for effects on cpu load? Currently I have simply stopped the old Rpi4 weewx service, being glad, that the new one works fine now. vince schrieb am Freitag, 15. November 2024 um 22:35:27 UTC+1: > In the absence of debug=1 logs, best of luck. > > As always, run with a 'clean' db and just one skin (pick one, Seasons or > Standard) to stabilize your installation with a known reference point > before adding your tweaks and additions and real db. That's the only way > we can try to replicate your setup if it comes to that. > > > On Friday, November 15, 2024 at 12:49:52 PM UTC-8 Vetti52 wrote: > >> My brief update: >> Today I have installed weewx on my new Rpi5 from the scratch. Then I >> copied the weewx.db from the old Rpi4. I also installed alltimeSeasons >> manually (which would be nice to be integrated into the Seasons skin more >> easily). Thus, although using historygenerator.py again, everything works >> fine, no detectable cpu load so far. I hesitate to install forecast, which >> uses it‘s own database. But I think, that was not the reason, because the >> 100% cpu happened without using forecast as well. >> Thus, it might remain an unresolved black box issue. >> >> BTW, at the Rpi4 journalctl only historygenerator.py is listed every 5 >> minutes. In the new Rpi5, there are now listed cheetahgenerator and >> imagegenerator as well. At least concerning the journalling there are >> differences between both setups. No idea, how this happened. >> >> Thanks anyway >> Peter >> >> vince schrieb am Sonntag, 18. August 2024 um 03:33:02 UTC+2: >> >>> Brief update - I had the user make his db available to me and it's fine >>> on both mac and raspi platforms, so the issue seems to be something in his >>> os+weewx setup..... >>> >>> On Saturday, August 17, 2024 at 9:22:18 AM UTC-7 vince wrote: >>> >>>> I don’t know. At this point all I can suggest is trying to determine if >>>> your db is the reason or not. >>>> >>>> If you can set up your second pi, install a new os and weewx using the >>>> Simulator only and get that working (without any uploaders enabled). Then >>>> scp your db over and restart weewx and see if it breaks anything. The >>>> initial run will take a few minutes to build the NOAA files but the >>>> archive >>>> runs should be very fast afterward. >>>> >>>> That will at least tell us if your db is ok. >>>> >>>> On Saturday, August 17, 2024 at 1:57:19 AM UTC-7 Vetti52 wrote: >>>> >>>>> Sorry, I did not mention earlier, but I checked the schema count long >>>>> time before, when changing to the new db schema. Yes, it is 114. >>>>> >>>>> But before I could check, I had do install sqlite3, which was not >>>>> installed in bookworm lite yet. weewx (or python) does not seem to need >>>>> this package by itself. >>>>> Then, before doing the apt update/upgrade, I had to remove the old >>>>> style gpg key and install the new one >>>>> >>>>> wget -qO - https://weewx.com/keys.html | sudo gpg --dearmor --output >>>>> /etc/apt/trusted.gpg.d/weewx.gpg >>>>> >>>>> Same with the bookworm reposities gpg key accordingly. >>>>> >>>>> Then I did, as mentioned in the other thread, check calc-missing only. >>>>> Dont know, if I should check all the others mentioned there, such as >>>>> appTemp. >>>>> >>>>> weectl database calc-missing >>>>> >>>>> and finally I checked for the amount of the 114 values, as you showed >>>>> above. However, I had to do it as >>>>> sudo echo ".schema" | sqlite3 /var/lib/weewx/weewx.sdb | wc -l >>>>> Maybe, there is a problem with my user rights settings? >>>>> >>>>> Tanks >>>>> Peter >>>>> vince schrieb am Freitag, 16. August 2024 um 20:41:53 UTC+2: >>>>> >>>>>> Check your weewx.sdb file first if it was originally on a system >>>>>> running weewx v3 or earlier. >>>>>> >>>>>> As the other thread and multiple other ones have indicated, there >>>>>> 'is' a change in v5 that calculates elements present in a skin that are >>>>>> not >>>>>> in the db schema. Typically this is appTemp but there are others. This >>>>>> typically occurs for sites that started in weewx v3 or earlier. New >>>>>> sites >>>>>> that started at v4 should not see this issue usually. >>>>>> >>>>>> Check your schema to see how many elements are in your actual db >>>>>> schema. The big wview-extended schema that is the default in v4 and >>>>>> later >>>>>> will return 114 for the following command. The original wview >>>>>> compatible >>>>>> schema used in v3 and earlier will report a number around 50 or so (I >>>>>> forget the actual number). >>>>>> >>>>>> # use the full path to your db below >>>>>> # or cd into the archive directory before >>>>>> # running the following command >>>>>> echo ".schema" | sqlite3 weewx.sdb | wc -l >>>>>> >>>>>> >>>>>> On Friday, August 16, 2024 at 6:49:32 AM UTC-7 Vetti52 wrote: >>>>>> >>>>>>> I have read the threads of Bartosz and those cited there but could >>>>>>> not find a solution for my problem there: >>>>>>> After updating to weewx 5.01. without any flaws after a few weeks, >>>>>>> when it got hot ouside, I realized, that my RPi4 had stopped working, >>>>>>> because of getting too hot. After restart everything worked fine, until >>>>>>> the >>>>>>> temperature again rised over more than 60 °C. I checked with s-tui that >>>>>>> one >>>>>>> of the four cores remained ad 100%. Finally I found, that as soon as >>>>>>> weewx >>>>>>> generated the first record, the cpu usage was 100%. Now, I moved the >>>>>>> RPi4 >>>>>>> into a cool and dark room and leave the RPi case open, so that the >>>>>>> temperature does not exceed 58 °C in average. >>>>>>> The latest update (and system upgrade) was yesterday. >>>>>>> My setup: >>>>>>> Raspberry4 with Bookworm, connected per LAN cable >>>>>>> Weewx 5.1 installed with Debian RPM >>>>>>> Ecowitt GW1000, driver 0.6.3 >>>>>>> Skin alltimeSeasons with forecast extension >>>>>>> >>>>>>> I replaced alltimeSeasons with regular Seasons skin, enabled debug = >>>>>>> 1 and restarted weewx. Still the same result: >>>>>>> >>>>>>> journalctl -u weewx >>>>>>> >>>>>>> Aug 16 14:20:20 raspbee weewxd.py[4438]: historygenerator.py: >>>>>>> Generated 5 tables in 0.20 seconds >>>>>>> Aug 16 14:22:49 raspbee systemd[1]: Stopping weewx.service - WeeWX... >>>>>>> Aug 16 14:22:50 raspbee systemd[1]: weewx.service: Deactivated >>>>>>> successfully. >>>>>>> Aug 16 14:22:50 raspbee systemd[1]: Stopped weewx.service - WeeWX. >>>>>>> Aug 16 14:22:50 raspbee systemd[1]: weewx.service: Consumed 1d 21h >>>>>>> 4min 31.966s CPU time. >>>>>>> Aug 16 14:23:09 raspbee systemd[1]: Started weewx.service - WeeWX. >>>>>>> Aug 16 14:23:16 raspbee weewxd.py[17471]: windy: version is 0.6 >>>>>>> Aug 16 14:23:16 raspbee weewxd.py[17471]: windy: Data will be >>>>>>> uploaded to https://stations.windy.com/pws/update >>>>>>> Aug 16 14:25:16 raspbee weewxd.py[17471]: historygenerator.py: >>>>>>> Generated 5 tables in 0.21 seconds >>>>>>> Aug 16 14:30:18 raspbee weewxd.py[17471]: historygenerator.py: >>>>>>> Generated 5 tables in 0.20 seconds >>>>>>> Aug 16 14:35:19 raspbee weewxd.py[17471]: historygenerator.py: >>>>>>> Generated 5 tables in 0.20 seconds >>>>>>> Aug 16 14:40:21 raspbee weewxd.py[17471]: historygenerator.py: >>>>>>> Generated 5 tables in 0.21 seconds >>>>>>> Aug 16 14:45:22 raspbee weewxd.py[17471]: historygenerator.py: >>>>>>> Generated 5 tables in 0.20 seconds >>>>>>> Aug 16 14:50:24 raspbee weewxd.py[17471]: historygenerator.py: >>>>>>> Generated 5 tables in 0.19 seconds >>>>>>> Aug 16 14:52:31 raspbee systemd[1]: Stopping weewx.service - WeeWX... >>>>>>> Aug 16 14:52:31 raspbee systemd[1]: weewx.service: Deactivated >>>>>>> successfully. >>>>>>> Aug 16 14:52:31 raspbee systemd[1]: Stopped weewx.service - WeeWX. >>>>>>> Aug 16 14:52:31 raspbee systemd[1]: weewx.service: Consumed 27min >>>>>>> 30.505s CPU time. >>>>>>> Aug 16 14:52:54 raspbee systemd[1]: Started weewx.service - WeeWX. >>>>>>> Aug 16 14:53:15 raspbee weewxd.py[18570]: windy: version is 0.6 >>>>>>> Aug 16 14:53:15 raspbee weewxd.py[18570]: windy: Data will be >>>>>>> uploaded to https://stations.windy.com/pws/update >>>>>>> Aug 16 14:55:15 raspbee weewxd.py[18570]: historygenerator.py: >>>>>>> Generated 5 tables in 0.20 seconds >>>>>>> Aug 16 14:57:33 raspbee systemd[1]: Stopping weewx.service - WeeWX... >>>>>>> Aug 16 14:57:33 raspbee systemd[1]: weewx.service: Deactivated >>>>>>> successfully. >>>>>>> Aug 16 14:57:33 raspbee systemd[1]: Stopped weewx.service - WeeWX. >>>>>>> Aug 16 14:57:33 raspbee systemd[1]: weewx.service: Consumed 2min >>>>>>> 22.493s CPU time. >>>>>>> Aug 16 14:57:33 raspbee systemd[1]: Started weewx.service - WeeWX. >>>>>>> Aug 16 14:57:40 raspbee weewxd.py[18714]: windy: version is 0.6 >>>>>>> Aug 16 14:57:40 raspbee weewxd.py[18714]: windy: Data will be >>>>>>> uploaded to https://stations.windy.com/pws/update >>>>>>> Aug 16 15:00:20 raspbee weewxd.py[18714]: historygenerator.py: >>>>>>> Generated 5 tables in 0.20 seconds >>>>>>> Aug 16 15:05:22 raspbee weewxd.py[18714]: historygenerator.py: >>>>>>> Generated 5 tables in 0.21 seconds >>>>>>> Aug 16 15:10:24 raspbee weewxd.py[18714]: historygenerator.py: >>>>>>> Generated 5 tables in 0.18 seconds >>>>>>> Aug 16 15:15:25 raspbee weewxd.py[18714]: historygenerator.py: >>>>>>> Generated 5 tables in 0.19 seconds >>>>>>> Aug 16 15:20:27 raspbee weewxd.py[18714]: historygenerator.py: >>>>>>> Generated 5 tables in 0.19 seconds >>>>>>> (No success to formate the pasted lines above into Courier, sorry) >>>>>>> >>>>>>> No idea, where else to look at now. My intention now is, to install >>>>>>> weewx on my other RPi5 from the scratch, and move the two databases >>>>>>> (weewx.db and forecast.db) over and reestablish alltimeSeasons and >>>>>>> forecast >>>>>>> afterwards. >>>>>>> So, is it possible to maintain weewx on different stations >>>>>>> simultaneously? Would my approach work like that? Or should I first >>>>>>> search >>>>>>> for the cause for 100% CPU load? >>>>>>> >>>>>> -- 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 [email protected]. To view this discussion visit https://groups.google.com/d/msgid/weewx-user/badfb385-3eaf-4389-b26e-7fccca203aa8n%40googlegroups.com.
