Funds stuck in AwaitingChannelReady for hours — cannot abandon/shutdown; restart only adds failure detail (reproduced twice with same peer)

[chukwuma619]

Summary

On fnn 0.8.1, a channel can stay in:

AwaitingChannelReady(OUR_CHANNEL_READY)

for hours after funding is confirmed on-chain, never transitioning to ChannelReady.

While stuck:

• abandon_channel is rejected (“please shutdown the channel instead”)
• shutdown_channel (cooperative) is rejected (invalid state)
• shutdown_channel(force=true) is rejected (invalid state)

I then restarted the node hoping it would recover. After restart, the channel was still stuck, and list_channels started showing:

failure_detail: "Peer disconnected during channel opening"

So that failure string is a post-restart symptom, not the original reason the open got stuck. Before restart, the channel was already stuck in AwaitingChannelReady with no recovery path.

Related: #1545, #606, #996.

Environment

• Network: CKB testnet
• fnn: 0.8.1

Timeline (current incident)

1. Opened a public channel with peer 03bf42…7b10 (I already had 3 Ready channels with this peer).
2. Funding confirmed on-chain.
3. Local state stayed in AwaitingChannelReady(OUR_CHANNEL_READY) and did not become Ready after waiting ~2+ hours.
4. Tried recovery RPCs while still in that state — all failed (see below).
5. Restarted fnn to see if state would heal.
6. After restart: still stuck in AwaitingChannelReady; now also shows failure_detail: "Peer disconnected during channel opening".
7. Peer is currently connected again (list_peers); reconnect + retry shutdown still fails.

Pattern I observed (hypothesis)

Same remote peer both times:

• Peer pubkey: 03bf4208b9ada97a5589ef9924819fe511b2a5e9ed3812df21a79435d39a147b10
• Peer multiaddr: /ip4/49.13.159.45/tcp/8228/p2p/Qmc9gzZB7aHYrVNxpakaQ8mLFchnnkesLhDfPPJEHQNbPt

In my cases this correlates with ~100k CKB aggregate capacity already open / opening with this peer (hypothesis only):

1. First wallet — large opens stuck (~100k + ~50k).
2. Second wallet — opened smaller channels bit-by-bit. First three became ChannelReady (~20k + ~30k + ~50k ≈ 99,703 CKB). A fourth ~50k channel then stuck the same way.

This is not “first channel with peer fails”; Ready channels with the peer already existed when the latest one got stuck.

Incident A — first wallet (earlier)

My node pubkey
03ea6d62f0dab7a9a4e1162f7f7a19b78db9d908468728ae5e2407713cbb09f0e6

Announced address
/ip4/197.211.63.22/tcp/8228/p2p/QmPWAqZUFxYwjH3b9VKT1P9Nh6cQMCbvqHUBxzqVjxkD8i

Stuck channels (fnn 0.8.1). Funding confirmed; both AwaitingChannelReady with OUR_CHANNEL_READY|CHANNEL_READY. After restart attempts, failure detail was also Peer disconnected during channel opening.

1. ~100k CKB

   • channel_id: 0x7cc479b706ce66e322c1c2dcacc2ca278e93d56c43eb31162a14046057457418
   • funding tx / outpoint: 0x852fac88c14602d51256511ecf7603706ac75fbb7f7219fab25377bc943f1dc1:0
   • created: 2026-07-10 01:30:15 UTC

2. ~50k CKB

   • channel_id: 0x39c4c7e014f77cccc837ebc1c52c4e3842b1bdd8df07d4154933da48224dbe6c
   • funding tx / outpoint: 0x92a701c9c394f69a2f69305c7dd96858f7957e8a2dbc5411e628764c44ae67ae:0
   • created: 2026-07-10 01:41:06 UTC

I then created a new wallet / new Fiber node and tried again with smaller opens.

Incident B — current wallet (reproduced)

My node pubkey
0340bd6865d097df1b8b9c423d37aadc215d4c0e582184f0ddeb71ff3cd8a875e5

Working channels with the same peer (all ChannelReady)

Capacity	channel_id	created (UTC)
19901 CKB	0x87dd4362ad8ee82d984d11c08df72e1c13e1667e450491c7a42a8d3f1855d334	2026-07-10 12:29:18
29901 CKB	0x67e32c9112afad1c21650846c5c5ae2b5012144f271b55017db61fa8e2a978b7	2026-07-10 12:30:45
49901 CKB	0x2b0333d6b9179a69aeaca576a7f847db71589fdde2fb8f80801aa9cbfca2c49c	2026-07-10 12:33:05

Ready total with peer ≈ 99,703 CKB.

Stuck channel

• channel_id: 0x897620a461ba4b80e6c3fa6ae4f30f909d9ea765737fc14a627eed2519404026
• state: AwaitingChannelReady / flags OUR_CHANNEL_READY|CHANNEL_READY
• failure (after node restart): Peer disconnected during channel opening
• local balance / capacity: 49901 CKB (remote 0)
• funding outpoint: 0xe66cc8817dad46a3af920192cf17eadb9211fa4d49c849815d6895838bde6d70:0
• funding tx (confirmed): https://testnet.explorer.nervos.org/transaction/0xe66cc8817dad46a3af920192cf17eadb9211fa4d49c849815d6895838bde6d70
• latest commitment tx hash: 0x09e2353ea3909e29637bc6d2b4aacb247d6a0e55f17f89f05e7fbfe59d905282
• created: 2026-07-10 12:37:30 UTC
• is_public: true, is_acceptor: false

RPC errors (while stuck / after restart)

abandon_channel:
Invalid parameter: Channel ... is in state AwaitingChannelReady(AwaitingChannelReadyFlags(OUR_CHANNEL_READY)), cannot be abandoned, please shutdown the channel instead

shutdown_channel (force=false):
Invalid state: Trying to send shutdown message while in invalid state AwaitingChannelReady(AwaitingChannelReadyFlags(OUR_CHANNEL_READY))

shutdown_channel (force=true):
Invalid state: Handling force shutdown command invalid state AwaitingChannelReady(AwaitingChannelReadyFlags(OUR_CHANNEL_READY))

Impact

• Funding is confirmed on-chain.
• Waiting hours does not complete ChannelReady.
• Shutdown/abandon are both blocked.
• Node restart does not recover funds; it only surfaces a disconnect failure detail.
• Result: CKB locked with no supported retrieval path from portable/production 0.8.1 RPC.

Happy to provide more logs if useful.