PC-BLE on RHEL 8.5

[A.P]

Hi!

We're working with a gateway comprised of a workstation running RHEL 8.5, a Bluetooth USB dongle with a nRF52840 chip running the connectivity_4.1.2_usb_with_s132_5.1.0 firmware. The gateway uses PC- BLE-DRIVER-JS to control the dongle and send/receive data.

Has anyone tried this before? Has anyone succeeded?

We're seeing repeating crashes of our process, and have yet to find the culprit for this.

Any help/ tips/ pointers to similar projects/ known issues or solutions would be appreciated.

Thanks!

[Jørgen Holmefjord]

Hi,

Maybe you can post the crashes/issues you are experiencing? That way it will be easier to give recommendations or compare towards existing issues.

As far as I know, we do not test pc-ble-driver on RHEL, we use Ubuntu for testing the Linux variant.

Best regards,
Jørgen

[Jørgen Holmefjord]

Can you try to increase the logLevel in adapter.open() to 'debug' or 'trace' and post the full log output here?

[A.P]

Yes. We're currently running on 'debug', and see no error. We've changed to 'trace' and will try to reproduce the issue.

I'll pass the info along to you as soon as I have something new

[A.P]

Another question in the meantime, if I may.

Would it be possible to ask Nordic to test the PC-BLE-DRIVER on Red Hat as well? We are willing to pay for this service, if needed. If this is relevant, please contact me directly.

Thank you!

[Jørgen Holmefjord]

Unfortunately, we do not do this kind of paid services. Adding support for another test platform would be a lot of work, but I can forward it to the developers as a suggestion to consider in the future.

[A.P]

Hi!

Ok, this is what we have found so far.
We're seeing this error bubbling up from the driver- Error: Error: [undefined] Error connecting to target device: Could not connect. Max number of connections reached..

For some reason, we've not seen this 'undefined' error pop up with months of testings on Debian workstations, but pops up after hours of running on Red-Hat workstations with the same environment.

Tweaking the way we handle this kind of error has resolved the apparent 'freeze'. So, it does not look like the driver actually 'froze' at any point, but that we had waited for an event which never happens after this error.

Hope this helps someone at some point, and if anyone could tell me why the difference between Debian workstations and Red-Hat workstations, I'd be grateful.