It's not Belchertown. It's not db size. I ran many tests. v5 and Belchertown can easily coexist on a pi3 or so system. If every time through you're exceeding the default archive period you're doing something custom to make that happen. I keep coming back to the known issue/slowness with synthetic measurements that Tom mentioned a few times recently.
Here's the test I just did with a newly flashed 32-bit pi3+ running a fully patched latest raspios... On a clean pi3+ with weewx v5 via pip and Belchertown installed, using the Simulator driver and no old pre-existing db present: Generated 8 files for report SeasonsReport in 3.45 seconds Generated 56 images for report SeasonsReport in 3.43 seconds Generated 12 files for report Belchertown in 6.93 seconds (elapsed time from writing to the db to Belchertown done was 14 seconds) I then stopped weewx, overwrote my weewx.sdb with my production system data with 17+ years of data, pre-seeded the Seasons NOAA tree so I didn't need to wait for them to be regenerated once for this test, then restarted weewx. So it's running the same skins with 17+ years of data, with the Seasons NOAA files having been pregenerated and dropped into place.... Generated 8 files for report SeasonsReport in 4.28 seconds Generated 44 images for report SeasonsReport in 6.75 seconds *Generated 236 files for report Belchertown in 266.59 seconds. <=====* (elapsed time from writing to db to Belchertown done was 4:37 so pretty close to a 5-minute default archive period) Why ? Because the 'first' Belchertown run generates its 'own' copy of the NOAA files down under the Belchertown public_html tree. It had to do that (for me) with 17+ years of month and year files which takes some time to complete. Running htop I saw two virtual cpus pegged at 100% throughout until it completed. After that Belchertown NOAA files are also pre-seeded there too so it's just doing normal incremental updates for future report runs. Each skin that generates NOAA files has to work hard once to do so vs. a pre-existing db if you start with a legacy db. You only have to bleed through that once. Subsequent runs should be quick. Lets see... So I let it run another cycle.....looks fine Generated 8 files for report SeasonsReport in 3.05 seconds Generated 18 images for report SeasonsReport in 1.22 seconds Generated 12 files for report Belchertown in 8.26 seconds (elapsed time this cycle was 14 seconds) And one more cycle.....looks fine Generated 8 files for report SeasonsReport in 2.94 seconds Generated 18 images for report SeasonsReport in 1.68 seconds Generated 12 files for report Belchertown in 7.80 seconds (elapsed time for this cycle was 14 seconds) Again - if you are continually exceeding perhaps 15 seconds for Belchertown to run on a pi3 or better you are doing something custom to make that happen. I just keep coming beck mentally to Tom's previous responses re: synthetic measurements. The original poster uses an Acurite which I don't know enough about to know for certain, but are you using that feature and not even knowing it ? Would providing your (sanitized) weewx.conf perhaps be of use in helping figure it out ? -- 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/9815dc0c-c883-43a3-b9e4-87f4e3a3986fn%40googlegroups.com.
