AddressV2 + authmailbox: incoming transfers only detected after a tapd restart

[cjc21]

Receiving a grouped asset I don't issue, via an AddressV2 with an authmailbox proof courier. Incoming transfers aren't picked up while tapd is running (tapcli addrs receives stays empty). After a tapd restart, the pending transfer is retrieved and credited. The next transfer is then missed again until another restart.

Asset: grouped, not issued by my node (synced from the shared universe)
Address: AddressV2, courier authmailbox+universerpc://universe.lightning.finance:10029

> tapcli universe federation config info

{
    "global_sync_configs": [
        {
            "proof_type": "PROOF_TYPE_ISSUANCE",
            "allow_sync_insert": false,
            "allow_sync_export": true
        },
        {
            "proof_type": "PROOF_TYPE_TRANSFER",
            "allow_sync_insert": false,
            "allow_sync_export": true
        }
    ],
    "asset_sync_configs": [
        {
            "id": {
                "group_key": "025d31a6603c1b6d4a21972f8c14a8f3833c6e86b8e81cce01b50b4d804d5f0180",
                "proof_type": "PROOF_TYPE_ISSUANCE"
            },
            "allow_sync_insert": true,
            "allow_sync_export": true
        }
    ]
}

025d31a6603c1b6d4a21972f8c14a8f3833c6e86b8e81cce01b50b4d804d5f0180 is the target group key.

So my question is:

• is authmailbox not working as expected?
• is something misconfigured on the universe side (allow_sync_insert for PROOF_TYPE_TRANSFER has to be true)?

[jtobin]

I think this is likely due to some missing TCP keepalive config on the authmailbox. I've added some in #2157, matching what we already use for price oracle connections.

[cjc21]

Thanks @jtobin . Will this make it into the next release?

[jtobin]

It did indeed! Let me know if v0.8.0 resolves your issue.