Well, it's been persistent for me. I noticed it months ago, when still running weewx 4.8 on Fedora 35 (still aarch64, aka 64-bit, version). I'm not sure how long it had been going on, I noticed that the Pi had OOMed it for using up all the RAM, before I limited it with systemd. But since I've upgraded to Fedora 37 (two major releases), that's basically everything getting rebuilt.
My attempts at finding python memory leaks in the past haven't been fruitful, but IIRC that was python2 (which I think didn't have tracemalloc). I'll see if I can poke at it. On Sunday, November 27, 2022 at 3:38:01 PM UTC-6 [email protected] wrote: > Every once in a while, a user will report the sudden appearance of a > memory leak. Then, just as mysteriously, the leak will go away. > > I can only assume that the cause is an underlying C library that got > "upgraded," then upgraded again. > > Do you recall the update history of your Fedora instance? Any recent > upgrades? Any involving Pillow? Database adapters? > > Python 3 offers the tracemalloc > <https://docs.python.org/3/library/tracemalloc.html> tool, which I've > found to be pretty good at finding the offending library. > > -tk > > > On Sun, Nov 27, 2022 at 1:10 PM Chris Adams <[email protected]> wrote: > >> I am running weewx 4.9.1 on Fedora 37 aarch64 on a Raspberry Pi 4. I have >> a Davis VantageVue weather system, connected via USB, and am uploading via >> FTP with SSL to a remote server. I don't have any plugins or anything else >> special installed/configured. >> >> The weewx python3 process grows at a pretty steady rate, a little over >> 150MB resident per minute (I have an update interval of 1 minute, so >> basically every run it grows). It's been showing a memory leak for a bit, >> previously with weewx 4.8 on Fedora 35 (I just waited until getting >> everything current before looking into it, hoping it'd resolve itself that >> way :) ). Right now, I'm running it under a systemd unit that restarts it >> every 6 hours (previously was capping the RSS memory to 192M, but that >> crashed the process with the OOM killer, so switched to scheduled restarts, >> which are "nicer"). >> >> There's nothing unusual in the logs, just the expected output like: >> >> Nov 27 15:07:16 python3[7451]: weewx[7451] INFO weewx.manager: Added >> record 2022-11-27 15:07:00 CST (1669583220) to database 'weewx.sdb' >> Nov 27 15:07:16 python3[7451]: weewx[7451] INFO weewx.manager: Added >> record 2022-11-27 15:07:00 CST (1669583220) to daily summary in 'weewx.sdb' >> Nov 27 15:07:19 python3[7451]: weewx[7451] INFO weewx.cheetahgenerator: >> Generated 8 files for report SeasonsReport in 2.34 seconds >> Nov 27 15:07:20 python3[7451]: weewx[7451] INFO weewx.imagegenerator: >> Generated 17 images for report SeasonsReport in 1.31 seconds >> Nov 27 15:07:20 python3[7451]: weewx[7451] INFO weewx.reportengine: >> Copied 0 files to /var/www/html/weewx >> Nov 27 15:07:25 python3[7451]: weewx[7451] INFO weewx.reportengine: >> ftpgenerator: Ftp'd 23 files in 4.71 seconds >> >> -- >> 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 on the web visit >> https://groups.google.com/d/msgid/weewx-user/a84edd0b-0f3f-4ea3-a246-5c15044d0c26n%40googlegroups.com >> >> <https://groups.google.com/d/msgid/weewx-user/a84edd0b-0f3f-4ea3-a246-5c15044d0c26n%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- 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 on the web visit https://groups.google.com/d/msgid/weewx-user/7b23c29a-5f91-4a8d-afc5-07af445b0a1dn%40googlegroups.com.
