Resource busy. BraunPWR BMS

[bialy39]

Hello.
I have very often this type of errors.
Is it means that somebody else is connecting to my BMS?

INFO [basebms] CER2601-022-030-020|D768: failed to initialize BMS connection (BleakDeviceNotFoundError)
ERROR [sampling] <bmslib.models.BLE_BMS_wrap.BMS object at 0x7f127dd70980> device not found, retry in 7 seconds (could not connect to 24:19:72:18:D7:68: [Errno 16] Resource busy: 'l2cap connect to 24:19:72:18:D7:68')
INFO [basebms] CER2601-022-030-020|D768: failed to initialize BMS connection (BleakDeviceNotFoundError)
ERROR [sampling] <bmslib.models.BLE_BMS_wrap.BMS object at 0x7f127dd70980> device not found, retry in 11 seconds (could not connect to 24:19:72:18:D7:68: [Errno 16] Resource busy: 'l2cap connect to 24:19:72:18:D7:68')
...
ERROR [basebms] CER2601-022-030-020|D768: TX BLE request error (BleakError)
...
INFO [bt] bt_diagnostics 24:19:72:18:D7:68: NOT seen during 3.0s scan on adapter=default (adapters=[{'index': 0, 'name': 'hci0', 'mac': '8A:88:4B:61:CA:44', 'bus': 'USB'}]) (26 other devices in range)
...
ERROR [main] 1 exceptions occurred fetching BMSs
ERROR [sampling] Error (num 1, max 200) reading BMS: TimeoutError at aiobmsble/basebms.py:435 in _await_msg()

Regards.

[fl4p]

No, most likely nobody over the air — it is something on your own Pi.

You are running ble_stack: bluek. That message is the kernel refusing the L2CAP connect with EBUSY, and there is only one place it can come from:

/* net/bluetooth/l2cap_core.c, l2cap_chan_connect() */
if (cid && __l2cap_get_chan_by_dcid(conn, cid)) {   /* cid = 0x0004, the ATT channel */
        hci_conn_drop(hcon);
        err = -EBUSY;

So the host already has a link to the BMS and something local already owns its ATT channel — normally bluetoothd, i.e. Home Assistant's own Bluetooth integration or another add-on that connected to the BMS. Unlike the BlueZ/D-Bus stack, which multiplexes one ATT channel between its clients, bluek needs CID 4 for itself, so it cannot share.

That also explains the other line: a peripheral stops advertising while it is connected, so the scan does not see it. One cause, two symptoms — not range, not a phone.

Check on the host:

bluetoothctl info 24:19:72:18:D7:68     # "Connected: yes" -> something local owns it
btmgmt con

Two fixes: keep HA's Bluetooth integration away from that MAC (ignore the discovered device), or set ble_stack: bleak so batmon and HA share bluetoothd's link.

I have also changed bluek: it now drops the offending link via mgmt before retrying, and reports BleakDeviceBusyError naming the real cause instead of "device not found". That will be in the next release.

[bialy39]

I'm running Home Assistant (HAOS) in Proxmox container.
I bought second battery and those errors are more often:

ERROR [basebms] CER2601-022-030-020|D768: TX BLE request error (BleakError)
ERROR [sampling] Error (num 1, max 200) reading BMS: TimeoutError at aiobmsble/basebms.py:435 in _await_msg()

Actually every 2-3 min.

When I run btmgmt con command I got Too small (0 bytes) get_connections rsp error
So I have no idea what to do.
I opened box and make photo of BMS.

BatMON recognizes it as BraunPWR.