-
-
Notifications
You must be signed in to change notification settings - Fork 699
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Startup if not all wallboxes, meters, integrations are reachable #14496
Comments
For inverter only (without an battery) there is the extra problem, that many of them go into a power saving mode after sunset and preventing evcc from restarting in the night, see #10972. |
Thanks for that description. I add this comment so that this issue can be found by users having that issue as well: |
I have similar issue with two wallboxes configured in evcc: as soon as one Wallbox is off or offline, then evcc cannot be used anymore for the other wallbox, even though it is online. evcc should be able to start in such case for the wallbox that is in operation. The Workaround to use two different yaml files is not a good WA. |
Ich musste leider gerade feststellen, dass dieses Verhalten sehr dringend notwendig wäre. Bei mir hat sich ein Shelly verabschiedet, der als PV Meter diente. (1 von 2 PV meters) Auch ein Failsafe Modus startete nicht. Eigentlich sollte der ja bereits vorhanden sein oder? Oder habe ich das falsch verstanden? |
Meine PV läuft seit 14 Jahren ohne einen einzigen Ausfall von irgendeiner Komponente. Eine Shelly ist auch eher Bastelware. |
Da stimme ich dir zu. |
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
@umrath @SolarPower2024 Lass uns nicht hier das Thema neu diskutieren. Konkrete Lösungen oder PRs sind natürlich immer willkommen. |
Wechselrichter im Powersafe-Mode macht gerade im Winter 16 Stunden am Tag permanent rote Fehlermeldungen in der UI. In meinem Fall startet evcc zum Glück trotzdem. Ich habe auch einen mobilen charger in evcc drinnen, wenn ich den an/abstecke muss ich derzeit die Config immer neu schreiben und evcc neu starten.
Das ist schön! |
Bei mir ist letztens das Internet ausgefallen und ich habe folgende Meldung im Log beim Start erhalten (Ausschnitt):
Entsprechend lief evcc nicht, obwohl Wallbox, etc. im lokalen Netz noch vollständig verfügbar waren. Generell versuche ich, möglichst unabhängig von Cloud-Diensten zu sein und verwende daher gerne Open-Source-Software und offene Komponenten. Mit dem Fehler ist es halt wieder sehr schade, an der Stelle vom Internet abhängig zu sein und das Auto nicht laden zu können. |
Leider ist dieser Issue überwiegend eine bewusste und so gewünschte Design-Entscheidung. EVCC funktioniert ja überwiegend problemlos auch ohne diese Daten. Ich bin am Überlegen einen Sidecar-Container mit einem HTTP-Proxy einzurichten, der einfach auf alle HTTP-Fehler mit einer gecachten 200 OK Response antwortet. Das wird bei einigen Modulen zwar zu Logik-Fehlern führen, aber EVCC kann sich in der Regel sehr gut davon erholen. Nur beim ersten Start werden alle Fehler als krtisch erachtet und der Start wird verweigert. In vielen Fällen kann man EVCC also nur in gewissen Situationen neu starten, bzw landet in einer gewissen Mieserie. Ich kann beispielsweise EVCC nur während die PV-Anlage Strom produziert starten, im Winter also nur zwischen 8 und 16 Uhr. Wenn ich also nach 16 Uhr EVCC stoppe, kann es sein dass es zu hohen Kosten kommt wenn dann Hausakku oder Auto bei extremen Strompreisen einfach weiterlädt. PS: evcc ist leider nicht open source im klassischen sinne (free software), sondern enthält viele kommerzielle Codestücke quer durch die code base, daher kann die community aus Lizenzgründen leider keine alternativen Versionen bereitstellen, die diese Fehler beheben. |
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
Hast du Home Assistant im Einsatz? Ich habe dies damit gelöst. Die Geräte sind in Home Assistant eingebunden und werden per mqtt an evcc geschickt. Sollten die Geräte wie zb mein PV oder grid meter nicht verfügbar sein, so sendet HA halt den Wert 0. |
Damit Tarifs kein Startup Error erzeugen gab es vor einiger Zeit #16258 |
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
Wenn es eine Design "entscheidung" wäre dann könnten wir hier als |
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
This comment was marked as off-topic.
As a reader of this discussion, I have perceived different wishes from various participants as to how the system should behave. I would like to contribute with a requirements proposal/clarification so that - assuming the proposal is accepted - the implementation concepts can go in this direction. Here is the proposal for how evcc should behave, initially regardless of the effort required to create this behaviour. I would like to simplify the system behaviour with the proposal by directing the start case (which is the issue here) directly to an operating state that occurs anyway. Here is the proposed requirement:
With requirement 6. evcc does not require its own boot procedure, but can be booted directly into the ‘all connections are down’ state. From then on, the recovery procedures for this state, which are required here anyway, take effect and, if all connected components are available, all connections are established shortly after the start - without the need for a separate boot procedure. I personally see the reason for these requirements in the assumption that the affected systems are all systems that are ideally in operation 24/7:
One indication that the systems should not wait for each other when starting up is shown in particular by the coupling of evcc with the wallbox: A wallbox is connected to the HEMS (evcc in this case) and vice versa. If both systems required the other when booting, a connection would never be established in first boot or after each complete shutdown, e.g. after power losses. I am fully aware that the behaviour described here cannot be implemented easily. With this proposal, I want to work towards a consensus on the desired behaviour. In other words: Is there a counter-reason why the boot behaviour should deviate from the behaviour when everything around evcc was once down in operation and is now available again? Personally, I can't see any reason right now, so I'd like to challenge that here. |
The reason right now is that the desired behavior is not in place. Nobody debates this is not being the desired state. |
My proposal introduces something new: it clearly states that no dedicated boot routine is wished. This is because there are different views on it in this thread, indicating that there are reasons for such a routine without clearly stating them.
As I said that is not true. There are serveral comments in this thread of people indicating that it is evcc's "design" to implement a fast fail routine at boot up. In fact, nobody stated the source of that requirement here but also nobody clearly contradicted that this requirement is not present. Examples: User:
User:
Reply:
(in other words: dear user, please fix your setup, rather than prioritizing a change in evcc's bootup behavior) The only comment indicating that the current boot behaviour is not wished it that one:
But also that one does not clearly state if and why a dedicated boot routine is required. If we agree on not having such a routine, the migration plan is (not easy) but clear:
This plan doesn't make sense if there are reasons for having a boot sequence. And those are indicated, but not stated so far. I don't know any, personally. That's why I'm challenging. |
Currently, evcc requires all configured devices to be reachable on startup. When this is not the case (e.g. a meter is not responding) evcc will run in a failsafe mode (control loop and many UI features disabled) and automatically restarts every 15min for retry until everything is reachable.
In some scenarios (unreliable hardware, broken devices), this behavior is problematic and requires temporary reconfiguration of evcc (e.g. disable a device) to get back into a functioning state.
To change this, we have to rework how initialization is done and may need to introduce concepts like device health status, per device retry mechanisms, discuss effects on e.g. load balancing, autodetect and visualize the affected parts correctly in the UI (e.g. disable loadpoint controls because a charger is not responding for a longer time).
\cc @andig
The text was updated successfully, but these errors were encountered: