We are currently attempting to switch our production environment to the GL-S20 [we typically use a Raspberry Pi with a dongle or an ESP-OTBR setup]. However, we are experiencing recurring stability issues with the GL-S20 (Firmware 2.0.0-R1). About once a week, the device requires a hard reset to restore functionality:
Nodes unable to send: Nodes lose the ability to transmit any data.
Missing ACKs (Most frequent): More oddly, nodes often send CoAP messages successfully, but they never receive the ACK back from the server.
Reviewing the logs, we see several entries that might point to the cause, though we are not certain:
Repeated RCP failure detected and P-RadioSpinel-: Trying to recover.
Frequent dropping radio frame: Parse and dropping radio frame: NoBufs.
Occasional Wait for response timeout warnings.
Is there a firmware version more stable than 2.0.0-R1?
Are there known limitations or specific configurations needed to support 40+ nodes without hitting NoBufs (buffer exhaustion) errors?
We have purchased 3 units via your B2C channel (GL.iNet Europe) for development and pilot testing.
Our requirements follow two distinct paths:
University Research Pilots: We deploy new testing installations every 2–3 months to evaluate new products and sensors. High stability is essential, as these units are critical for our data collection.
Commercial Integration (Spin-off): The GL-S20 is currently the recommended gateway for a certified commercial product with an expected rollout in 2027. For these third-party installations, the responsible team will engage with your B2B department to establish a formal relationship.
We are available to perform tests across different setups or firmware versions and can provide detailed OTBR logs through your specified technical support channel.
Please contact our sales representative via email at sales@gl-inet.com, including the link to this post. We will then make further arrangements based on your project details.