Should the last command be: - weectl database calc-missing Sent from my iPad. On 30 May 2024, at 03:28, Mark Jenks <mjenks1...@gmail.com> wrote:
I ran this, and did the recalc-missing, and it seems to be running better. Except there is only one record in those tables. Not sure why that would cause that issue. weectl database add-column appTemp weectl database add-column cloudbase weectl database add-column visibility weectl database add-column windrun weectl database add-column cloud_cover weectl database add-column aqi weectl calc-missing On Tuesday, May 28, 2024 at 8:10:45 PM UTC-5 Mark Jenks wrote:
Seasons itself doesn't do it, and Belchertown with no settings on it is doing it. So, I'm adding those columns now and recalculating now.
I think this was the Problem with missing Database extended schema object. I had the same problem as i upgraded from weewx 4 to 5. I solved this with.
weectl database add-column appTemp weectl database add-column cloudbase weectl database add-column visibility weectl database add-column windrun weectl database add-column cloud_cover weectl database add-column aqi weectl calc-missing weectl calc-missing appTemp
vince schrieb am Dienstag, 28. Mai 2024 um 19:38:07 UTC+2:
The other test would be to disable belchertown, enable seasons, and see if that’s ok. That would check the db vs. the v5 calculated aka synthetic readings feature that slowed some folks down.
On Tuesday, May 28, 2024 at 10:25:10 AM UTC-7 Tom Keffer wrote:
I thought of the appTemp angle, but those queries involve thousands of small aggregate intervals, not one query over 4 years.
Still, it's worth a try.
You seem to be hacking randomly and praying. I'd suggest deleting Belchertown at this point, verifying it is deleted, then add it back in. Do not customize it yet. Just run the default setup, it works. Really.
My guess is you are missing the appTemp element in the db. There are a bunch of threads on this re: Belchertown and v5 weewx.
Check your mariadb schema and see if it is missing appTemp. If you have the old wview-compatible db you'd have about 50 elements in it. If you later switched to the v4 and above bigger wview-extended schema, you'd have 114 elements in it. Sorry but I don't know the mariadb commands to count them nor check.
For sqlite3 it would be: 'echo ".schema" | sqlite3 mydbname | wc -l' On Tuesday, May 28, 2024 at 3:59:19 AM UTC-7 Mark Jenks wrote:
I removed the last two from skin.conf, and it still does the large query. I'll keep digging tonight.
generator_list = weewx.cheetahgenerator.CheetahGenerator, weewx.reportengine.CopyGenerator, user.belchertown.HighchartsJsonGenerator
On Monday, May 27, 2024 at 10:06:02 PM UTC-5 Mark Jenks wrote:
false alarm so far. mqtt is tied to belchertown. So, seems to be belchertown like you suspected.
Digging more tomorrow.
On Monday, May 27, 2024 at 10:00:15 PM UTC-5 Mark Jenks wrote:
I just ripped mqtt out also. I'll check it in the morning and see if that makes a difference.
On Monday, May 27, 2024 at 9:52:21 PM UTC-5 Mark Jenks wrote:
Bad news? I uninstalled belchertown, and it's still hitting high CPU on mariadb and python3. This is the hit to the DB.
SELECT * FROM archive WHERE dateTime > 1704088800 AND dateTime <= 1735711200 ORDER BY dateTime ASC
On Monday, May 27, 2024 at 7:16:36 PM UTC-5 Tom Keffer wrote:
That query is asking for every single record over 4 years of data --- about 400,000+ records with a 5 minute archive interval. That's not the use pattern when an xtype is causing the problem.
It's hard to imagine why the Belchertown skin would need data at that density. Try isolating the problem by shutting off the imagegenerator, then the cheetahgenerator (you can do that by modifying generator_list in skin.conf).
Then once you know which one is the culprit, then start trimming their respective sections in skin.conf until you isolate the plot or tag that is causing the problem.
Just built a new weewx on my Fedora 38 this morning, and attached it to my mariadb. It has 13 years worth of data in it.
I am running Belchertown and mqtt, and installed Windy. But I just removed Windy to see if that was it, but still does it. Python and/or moriadb goes to 100% CPU. Mariadb says the high query during that time is: SELECT * FROM archive WHERE dateTime > 1320346500 AND dateTime <= 1716824400 ORDER BY dateTime ASC But that makes no sense, since I can query my entire archive table (select *) in about 3 seconds.
Any thoughts on how to catch what is going on? I could enable debug, but hoping for a better way other that looking at a ton of logs.
Thanks!
--
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+...@googlegroups.com.
--
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/ae663ace-fe2d-4951-be0f-2fbe54c57944n%40googlegroups.com.
--
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/C3A4FF4E-B352-46B2-B7FC-AC7023183389%40btinternet.com.
|