search

espressif / esp-aws-iot

AWS IoT SDK for ESP32 based chipsets
Apache License 2.0
256 stars 154 forks source link

OTA Job shows status Failed even after ESP32 successful update (CA-240) #137

Open law-ko opened 1 year ago

law-ko commented 1 year ago

Hello,

We have tested the ota_mqtt example and everything works fine except the Status showed failed even after ESP32 has successfully updated via OTA.

Are there any messages we need to send back to AWS IoT to confirm the update is successful and complete?

Screen Shot 2022-09-19 at 11 22 43 PM

Thank you.

law-ko commented 1 year ago

Any updates on this?

davidallenmann commented 1 year ago

I got the same error until I increased the APP_VERSION_BUILD in demo_config.h. Used IDF v4.4.2

SolidStateLEDLighting commented 1 year ago

In theory, once you have rebooted, the system should go check for any in-progress jobs. It will find the uncompleted OTA that you previously were running. Then, your software would essentially resume that job and send the final update back to AWS which will render that job as successfully complete. I don't use those examples --- so, I don't know how it is technically done in those example projects.

bhagirathsinh-vaghela commented 10 months ago

Possible solution: I might be too late and you may have figured it out. I saw the same and observed that it is dependent on CONFIG_BOOTLOADER_APP_ROLLBACK_ENABLE.

I am not totally sure if my understanding is correct(I don't work on c projects). But this is what I think is happening... If this is not enabled(not CONFIG_BOOTLOADER_APP_ROLLBACK_ENABLE=y in sdkconfig), after the OTA job execution is done and new firmware is loaded to one of ota partition, the otadata on the otapartition needs to get set as ESP_OTA_IMG_NEW but it gets set as ESP_OTA_IMG_UNDEFINED instead. The very first boot of the new firmware should see this and try to set new otadata as ESP_OTA_IMG_PENDING_VERIFY but that doesn't happen and ultimately we never get to setting the otadata as ESP_OTA_IMG_VALID and thus the OTA execution status is not updated as successful.

In my case, I was seeing that second reboot happens which sees the running app version as the same as what AWS OTA job document says and thus marks the job execution as FAILED with "reason":"rejected: 0x00000010"(log said: New image is being rejected: Application version of the new image is invalid: OtaErr_t=OtaErrSameFirmwareVersion). I hope this helps.