If you're running a throwaway vm, try running the benchmarks at
https://github.com/weewx/weewx/wiki/Benchmarks-of-file-and-image-generation
and see what your numbers look like. You should see numbers well under 3
seconds on that kind of gear.
On Tuesday, December 26, 2023 at 6:52:42 AM UTC-8
I ran some more tests with my main DB converted to sqlite, using weectl to
run repots using the default Seasons skin.
The times remained basically constant at about 4.5 minutes user+sys CPU
as I kept deleting old records.
As I got below 500k records the times dropped proportionally to the
The wmr300 DB has 3.7million records, and the ecowitt DB has only 530k.
The system is a VM running on an i5 - the VM is allocated 8GB ram and 4
cores. It was not stressed at all by weewx V4.
The CPU usage is all in the python, not in the mariadb server.
While the python code is thrashing
Hey All,
I’m going to give this a go, but want to confirm based on this:
https://weewx.com/docs/5.0/upgrade/
I’m just running on a raspi, and installed via apt, so I should just be able to
do as documented here and upgrade this way? I swear a few months back there
wasn’t going to be a
Hey Tom,
Thanks! Should I assume I should stop the current weewx, backup the db, etc?
And yup, realize it’s a temp repo. =)
Thanks!
From: Tom Keffer
Sent: Tuesday, December 26, 2023 3:20 PM
To: rcsta...@gmail.com
Cc: weewx-development
Subject: Re: [weewx-development] V5.0 release
So, it sounds like you can't reproduce it either...?
Because it looks like a hard crash of the Python interpreter, it is likely
it was caused by a bad cache, or something like that.
On Tue, Dec 26, 2023 at 1:46 PM Vince Skahan wrote:
> What I did was install a clean beta via pip, then
Can you please run the benchmark that Vince linked to earlier:
https://github.com/weewx/weewx/wiki/Benchmarks-of-file-and-image-generation
Unfortunately, it runs only under V5 (not v4.10). Still, it will give us an
idea whether you're in the right ballpark for runtimes.
-tk
On Tue, Dec 26, 2023
That time may be skewed by my prod weewx being in it's current state.
Stopped the production instance and re-ran:
weectl report run --config=./weewx.conf
Using configuration file *./weewx.conf*
All enabled reports will be run.
Generating as of last timestamp in the database.
2023-12-26
I copied the benchmark DB and the conf file onto my main weewx
installation, edited the conf file accordingly, and got the same benchmark
results, about 1.5 seconds each.
So I can only assume there is something in my customisations that V5 is
having trouble with.
On Wednesday 27 December
I am trying to reproduce the result that Vince and Cameron have reported,
but am unable to do so. More details would be helpful.
Did you both do an upgrade from v4.10.2 to v5.0.0rc1?
Did you have any extensions installed?
Vince, I know you were running "weectl report run" using Python 3.11,
What I did was install a clean beta via pip, then overwrite extension.py
down buried in the venv from the later one you'd uploaded to github. I
'think' that I hand-edited in one patch as you'd previously updated
multiple things after the beta was released, but my memory is a little hazy
Yes, there is a packaged version, so, yes, you do a package upgrade. NOTE:
the release candidate is in a temporary apt repository, as outlined in the
first message of this thread.
On Tue, Dec 26, 2023 at 3:11 PM wrote:
> Hey All,
>
>
>
> I’m going to give this a go, but want to confirm based on
Mine is an upgrade from 4.10.2, and probably all the way from the original
v4, at which stage I think is was a clean apt install from V3 and I think
that was when I switched to python 3
.
My system is up-to-date Deb 12, so Python 3.11.2-1+b1
I can reproduce it by simply removing the symlink
Hey All,
Okay, stopped weewx, did the change in sources, and an apt upgrade. Declined to
replace weewx.conf, and then updated the existing one to have the weewx root
properly, and start weewx. Started right away. Grabbed the one archive record
that it missed during my process.
Files
In my case, benchmark:
weectl report run --config=./weewx.conf
Using configuration file *./weewx.conf*
All enabled reports will be run.
Generating as of last timestamp in the database.
2023-12-26 18:42:45 weectl[1787]: INFO weewx.cheetahgenerator: Generated 21
files for report SeasonsReport
ok that looks like the old bug where extensions were in one place and some
of weewx was looking in other places (at one point in the betas)
On Tuesday, December 26, 2023 at 6:03:17 PM UTC-8 Cameron D wrote:
> Mine is an upgrade from 4.10.2, and probably all the way from the original
> v4, at
74 seconds clock time to run a seasons report looks like the same problem I
am seeing.
On Wednesday 27 December 2023 at 11:25:17 am UTC+10 rcst...@gmail.com wrote:
Hey All,
...
Here’s 4.10
Dec 26 17:00:25 raspi-server-misc weewx[25329] INFO weewx.cheetahgenerator:
Generated 8 files for
Benchmark results: 1.4s for files and 1.5s for images
Added to Wiki.
On Wednesday 27 December 2023 at 12:02:02 pm UTC+10 Tom Keffer wrote:
> Can you please run the benchmark that Vince linked to earlier:
> https://github.com/weewx/weewx/wiki/Benchmarks-of-file-and-image-generation
>
>
Clean pi5 64bit installation of the beta dpkg. I manually copied my prod
weewx.sdb (17+ years of data) and its previously generated NOAA files into
place before starting weewx.
Dec 26 18:20:15 raspberrypi weewxd[12806]: INFO weewx.manager: Added record
2023-12-26 18:20:00 PST (1703643600) to
Ran this again, realizing I hadn't removed public_html between first and
second run (thinking it was production weewx causing the issue).
root@raspi-server-misc:/var/tmp/weewx-benchmark# weectl report run
--config=./weewx.conf
Using configuration file *./weewx.conf*
All enabled reports will be
I just wiped weewx from my test system and did *a fresh install* - I got
5.0.0-rc2.
There was no weewx.conf created, so I ran dpkg-reconfigure and there is
still no weewx.conf.
I added back the benchmark conf and DB. On trying the manual report gen it
exited immediately, and syslog showed
21 matches
Mail list logo