Closed odlg closed 10 months ago
Please share the full log. There is something odd with the start timestamps:
plan: charge 0s (required: 1h16m5s) starting at 0001-01-01 00:53:28 +0053 LMT until 2023-11-03 10:00:00 +0100 CET (power: 3680W, avg cost: NaN)
We need to find out where/why the start timestamp goes wrong.
Im Log wird leider nicht klar, wo die komische Startzeit her kommt. Es scheint, als hätte der Planner einen Plan ohne Inhalt gemacht- da würde auch zu den NaN Kosten passen. Warum das passiert ist leider absolut nicht erkennbar. Falls Du das Szenario reproduzieren kannst würde ich mich über Beschreibung freuen. As-is gibts leider nix was wir tun könnten :/
@andig I have not been able to not reproduce this. How do I debug this to actually make it work? I have tried to mount /etc/localtime from the docker host, also mounting /usr/share/zoneinfo/Europe/Berlin to match the TZ env. It does not change anything.
It does not change anything.
Change what?
How do I debug this to actually make it work?
Now you got me confused. To make what work? Are you saying Planner never works for you? That would be enough of a reproducer! In that case- please add steps to reproduce.
Planner never worked for me. The steps to reproduce:
@andig I just tried to run 0.122.0 directly on the host OS (so not in Docker) and it has the same issue with Planner not working. It is also logging the weird LMT timezone.
Mhhm. I cant repro. Are you able to see that in the demo instance or using the demo.yaml config? I need something to attach the debugger to…
@andig demo.yaml does not reproduce the issue. That logs just fine charge plans.
[lp-1 ] DEBUG 2023/11/05 20:07:07 plan: charge 2h52m52s (required: 3h36m49s) starting at 2023-11-05 20:07:08 +0100 CET until 2023-11-05 23:00:00 +0100 CET (power: 3680W, avg cost: 206.976)
@andig Using demo.yaml with tariffs from my config above triggers the issue.
[lp-1 ] DEBUG 2023/11/05 20:21:39 plan: charge 0s (required: 3h41m36s) starting at 0001-01-01 00:53:28 +0053 LMT until 2023-11-05 23:00:00 +0100 CET (power: 3680W, avg cost: NaN)
Please provide exact steps. Which loadpoint, vehicle, limit etc. Still cant repro.
@andig Using evcc.yaml and no evcc.db file I did these steps:
Set a charge plan for 'blauer e-Golf' to a few hours from now. Immediately it logs
[lp-1 ] DEBUG 2023/11/05 21:29:20 plan: charge 0s (required: 3h39m12s) starting at 0001-01-01 00:53:28 +0053 LMT until 2023-11-05 23:00:00 +0100 CET (power: 3680W, avg cost: NaN)
Could you please try 0.122.0? It has some additional logging I think. Doing as you describe (and deleting evcc.db) I get:
go run main.go -c cmd/demo.yaml
[main ] INFO 2023/11/06 20:17:53 using config file: cmd/demo.yaml
...
[lp-1 ] DEBUG 2023/11/06 20:18:12 set target time: 2023-11-06 23:00:00 +0100 CET
...
[lp-1 ] DEBUG 2023/11/06 20:18:18 plan: charge 2h41m42s (required: 3h38m24s) starting at 2023-11-06 20:18:18 +0100 CET until 2023-11-06 23:00:00 +0100 CET (power: 3680W, avg cost: 172.000)
Are you able to repro with the demo.yaml? Unfortunately I can't use your yaml sind the devices are different :(
With the orignal demo.yaml i do not reproduce the issue. With https://github.com/evcc-io/evcc/files/13260725/evcc.yaml.txt I do reproduce the issue - it is demo.yaml with energinet tariffs
Great find, thank you! It turns out, that Energinet violates a core (and sane) assumption of the planner
// rates are by default sorted by date, oldest to newest
last := rates[len(rates)-1].End
since it's API returns time slots in reverse order (why would anybody do this?). PR ensures all tariffs are sorted in ascending order now, just to be safe.
I have rebuilt the Docker image, it looks good - now I hope my car is charged tomorrow :-)
Describe the bug
A planned charge does not start.
Steps to reproduce
In Solar mode, set a target limit higher that current charge, set departure time. Setting Smart Grid price limit has no effect. Running in docker on x86_64 Linux.
Configuration details
Log details
What type of operating system are you running?
Docker container
Version
0.121.5