Change many devices in one pass

Bulk change in the header, in the selection bar or under the arrow next to Change config in the details of one device opens the same dialog. One device or forty, the same rules apply. Nothing is sent to a board before you press Create jobs.

Bulk change covers the basics: firmware, interval, the WiFi signal and the power mode. To change everything on one device, its sensors, pins, options and how often each value is sent, press Change config in its details. The device builder opens with the saved setup, the board stays fixed, and the last step sends the new setup over the air. A board on firmware 0.5.0 or later takes new sensors, other pins or another sleep mode this way without a firmware update. A sensor that needs a newer firmware says so in the builder, and the update goes first in the same job.

  1. 1
    Change
    Tick what changes and pick the new value. Choose the verify mode. Firmware and config combine into one pass.
  2. 2
    Devices
    Tick the devices, or use Quick pick: All from a template, All with update available, or Clear. From 20 devices on, a filter field helps.
  3. 3
    Check & review
    HydroNode checks every device and shows what changes where. Create jobs starts the change for every device that can take it; the others are skipped.
ChangeWhat it does
FirmwareInstalls a signed release. ESP32 family only. The latest is preselected; withdrawn releases are not offered.
Send intervalHow often the board reports: 1, 2, 5, 10, 15, 30 or 60 minutes.
Add valueSends the WiFi signal (RSSI in dBm) with every reading. WiFi boards only.
Power modeAlways on, Modem sleep, Light sleep, Deep sleep or Hibernate, as far as the chip supports it. A mode none of the picked devices can run is greyed out with the reason, for example SGP30 needs always-on.

Reading the check

The check runs every rule for every device before anything is created. The box Same for every device shows the change most devices get, for example Send interval 10 min → 5 min. Below it, one row per device and one cell per change. A cell is yellow when a device starts from a different value than most, for example differs from the usual 10 min.

ResultMeaning
OKThe device takes the change as shown.
WarningThe device takes the change, but read the reason first. Skipped parts, low battery and a silent device land here.
Not possibleNothing is sent to this device. The reason stands in the row, which is dimmed.
The check saysWhy
Needs one USB reflash to 0.5.0The firmware on the board cannot update itself yet.
ESP8266 takes config onlyFirmware is skipped on an ESP8266. Config parts still go.
Not HydroNode firmwareOwn sketches and ESPHome are not updated by HydroNode.
No saved setup, reflash once with the builderConfig changes start from the saved setup.
EC probe needs always-onA sensor that samples all the time cannot sleep deep. The name comes from your setup.
Outputs need the board awakeRelays and LEDs cannot keep their state in deep sleep.
Deep sleep needs at least 30 sEach power mode has a shortest interval.
A job is still runningOne open job per device. Wait for it or cancel it.
Not reporting, the job waits for its next contactThe device is delayed or offline. The job starts when it is back.
Already reports sensor read errors, Strict may roll backA sensor failed in the last 24 hours. Fix it, or use Lenient for this change.
Changed by hand since the template, your edits are replacedA template apply overwrites the drifted setup.
The configuration is too large for the device.The new config does not fit into the 8 KB config area of the board.
The board cannot change over the airChange config keeps the board. For another board, flash it with the device builder.
SGP40 needs firmware 0.5.2 or newerThe sensor came with a later firmware. Update the firmware in the same change, or over USB on an ESP8266.
Below 30 % battery you confirm twice. A firmware download is the most power hungry thing a board does. The first click on Create jobs names the weak devices and turns the button amber. The second click creates the jobs. If the battery dies during the update, the old firmware starts again, but the board may stay quiet until it is charged.

From Queued to Installed

Every device in a change gets one job. The details panel draws it as five steps: Queued, Offered, Download, Verify, Installed. The board always starts the step itself, when it checks in. HydroNode never reaches into your network.

  1. 1
    Queued
    The job exists and waits for the board to check in. A sleeping board shows Waiting for check-in with the time of its next round.
  2. 2
    Offered
    The reply to the next reading carries the update. The board finishes its round first, so no reading is lost.
  3. 3
    Download
    The board checks the signature, then loads the firmware into its free slot and verifies its checksum. An interrupted download resumes where it stopped.
  4. 4
    Verify
    The board restarts into the new firmware and checks itself in its first wake cycle. The check after the restart is described in the OTA guide.
  5. 5
    Installed
    The check passed. The details show for example "Verified Strict in 41 s", and the saved setup follows the new version.

A config change takes the same path without a download: the board receives the new settings with the reply, keeps its old config, writes the new one, restarts and checks. A change with firmware and config runs the firmware first and the config right after, as one job.

WhenWhat happens
The board is asleepThe job waits for its next round. Nothing wakes a sleeping board early.
Offered three times, not takenThe job fails with "Offered 3 times, not taken". Check the console for refused offers.
The download stalls for 10 minutesThe job goes back to Offered, and the next reply offers it again.
HydroNode was unreachable during the checkThe board rolls back, the job is queued again and tried in a later round, up to five times.
No report after the restartAfter three intervals plus five minutes without a word, the job fails with "No report after restart".
The device changes owner or a share endsJobs the board has not taken yet end when the person who started them no longer manages the device.

Cancel a job

While a job is Queued, Offered or Waiting for check-in, a Cancel button stands next to its state. Cancelling keeps the board exactly as it is: nothing was downloaded yet. Once the download or the config change started, the job runs to its end, and a failed check rolls it back on its own.

Bulk change can also set the device thresholds of HTTP devices. A LoRaWAN station takes its interval and thresholds with Change config in its drawer, one station at a time: Battery and Power.

What the board does during an update, the check after the restart and the rollback are explained in Over-the-air updates.

NextTemplates