Moving a user to another voice channel makes them invisible and stuck reconnecting

(#2977) Bug Open: Confirmed Priority: P1 desktop moderation voice

Reproduction steps

  1. Move User

Observed behaviour

The user gets moved to the channel, but is invisible, and gets disconnected and reconnected constantly in a loop

Build information

Canary Desktop 2026.606.140333, Web 2026.605.11941, Windows 11 10.0.26300 (x64), Electron 41.2.2, Chrome 146.0.7680.188, Node 24.14.1, Locale en-US

Platform

Windows

Details

Expected behaviour Should move the user to the channel without issues
  1. Rex changed the status from Needs triage to Confirmed

Merged posts

These posts were merged into this one. Their comments are now part of the conversation below, marked with where they came from.

Report details

Summary

When any user (and bots) with Move Members permission moves me to another voice channel, the Flatpak+AppImage client disconnects from VC entirely instead of reconnecting to the new channel. The web version handles this correctly. This is reproducible with any server-initiated move by a user.

Steps to reproduce

  1. Join a voice channel on the Flatpak or AppImage client
  2. Have another user with Move Members permission drag you to a different voice channel
  3. Client disconnects entirely instead of joining the new channel

Environment (optional)

  • OS: KDE Plasma + CachyOS
  • Client: Fluxer Desktop (Flatpak build + AppImage)
  • Web Version (web.fluxer.app): works correctly

Logs or screenshots (optional)

Logger.tsx:131 [11:42:27] [VoiceConnectionManager] [Info] VOICE_SERVER_UPDATE: server-initiated channel move {expectedChannelId: '1484051274592260409', incomingChannelId: '1484051529215836948', connectionId: 'nilgiri-barometric'}
Logger.tsx:131 [11:42:27] [VoiceSubscriptionManager] [Info] All subscriptions cleaned up
Logger.tsx:131 [11:42:27] [MediaEngineFacade] [Info] All tracking stopped
Logger.tsx:131 [11:42:27] [VoiceConnectionManager] [Info] Attempting to connect to LiveKit {endpoint: 'wss://601565d.nue.fluxer.media', guildId: '1484051274592260405', channelId: '1484051529215836948'}
console.js:36 disconnect from room {room: 'guild_1484051274592260405_channel_1484051274592260409', roomID: 'RM_7jHyzaqLzc2B', participant: 'user_1471908307644801048_nilgiri-barometric', pID: 'PA_FXjWPBi5SfSu'}
Logger.tsx:131 [11:42:28] [VoiceParticipantManager] [Info] Hydrated participants {count: 1}
Logger.tsx:131 [11:42:28] [VoicePermissionManager] [Info] Complete {deafened: false, participantCount: 0}
Logger.tsx:131 [11:42:28] [VoiceConnectionManager] [Info] Connection established
Logger.tsx:131 [11:42:28] [RouterUtils] [Info] transitionTo /channels/1484051274592260405/1484051529215836948
Logger.tsx:131 [11:42:28] [VoiceSubscriptionManager] [Info] All subscriptions cleaned up
Logger.tsx:131 [11:42:28] [VoiceSubscriptionManager] [Info] Room set {participantCount: 0}
Logger.tsx:131 [11:42:28] [VoicePermissionManager] [Info] Complete {participantCount: 0}
Logger.tsx:131 [11:42:28] [MediaEngineFacade] [Info] All tracking started
Logger.tsx:131 [11:42:28] [VoiceConnectionManager] [Info] LiveKit connection succeeded
Logger.tsx:131 [11:42:28] [VoiceConnectionManager] [Info] Initializing voice connection
console.js:36 publishing track {room: 'guild_1484051274592260405_channel_1484051529215836948', roomID: 'RM_rKBRuhcZUgFJ', participant: 'user_1471908307644801048_nilgiri-barometric', pID: 'PA_JrGMZPN4xSpd', trackID: undefined, …}
Logger.tsx:131 [11:42:28] [VoiceSubscriptionManager] [Info] All subscriptions cleaned up
Logger.tsx:131 [11:42:28] [MediaEngineFacade] [Info] All tracking stopped
Logger.tsx:131 [11:42:28] [VoiceConnectionManager] [Info] Connection terminated {reason: 'error'}
Logger.tsx:131 [11:42:28] [VoiceMediaManager] [Info] Successfully enabled microphone
Logger.tsx:131 [11:42:28] [VoicePermissionManager] [Info] Started permission watching {guildId: '1484051274592260405', channelId: '1484051529215836948'}
The failure occurs ~1 second after [VoiceConnectionManager] Initializing voice connection. Subscriptions and tracking are immediately cleaned up, followed by Connection terminated {reason: 'error'}. A stray publishing track then fires with trackID: undefined on a different roomID. Tested flatpak run --device=all and that does not fix the issue. The web client handles the same aVOICE_SERVER_UPDATE flow without any errors.
Report details

Reproduction steps

  1. Join a voice channel.
  2. As a moderator, click on the user's profile and move them to another voice channel.

Observed behaviour

In Fluxer mobile, the user doesn't properly enter the channel. And when they do enter, it doesn't appear that they are on a voice call, Neither for him, nor even for other Fluxer users.

Build information

Version 2026.09.100, Beta 3, Channel Beta, Build 2530, Android 12 (arm64), Device samsung SM-A127M, Locale pt-BR, Push fcm, Flutter 3.47.5, Dart 3.13.4

Platform

Android

Details

Expected behaviour Move the user to another voice channel, and make it appear that they are on the call, not becoming "invisible" to other users.
Report details

Reproduction steps

  1. User A joins voice channel 1
  2. User B (with Move Members permission) drags User A to voice channel 2
  3. User A disappears from channel 1's member list and appears in channel 2
  4. Users in channel 1 can still hear User A speaking
  5. User A's audio is not heard in channel 2

Observed behaviour

When a user is dragged (moved) from one voice channel to another by a moderator, the user disappears from the channel list in the UI but their audio is still heard by other participants in the original channel. The moved user appears to be in the destination channel visually, but their voice stream is not properly transferred — it remains connected to the original channel's LiveKit room.
  • UI updates correctly (user moves in the list)
  • Audio stream stays connected to the original channel
  • Creates a "ghost audio" situation
Expected behaviour User A should be fully moved to channel 2: audio transfers to the new channel and is no longer heard in the old one.

Details

Environment
  • Fluxer Canary (web app and desktop)
  • Hosted instance (api.fluxer.app)
Additional context
This is different from #296 and #715 (which are about the API PATCH endpoint disconnecting users entirely) and #434 (Flatpak/AppImage disconnecting on move). This is a client-side desync where the UI updates but the audio connection doesn't follow.
Report details

Reproduction steps

  1. Join a voice channel.
  2. Have an admin move you to a different voice channel.
  3. See that the user has vanished from the voice channels in the UI.
  4. Join the channel they were moved to.
  5. Listen to them connect (they will appear in the large right window, but not in the channel list still)
NOTE. API checks ALSO don't see the user in the channel they were moved to (I found this while testing a bot that moves users between channels, have since reproduced without the bot)

Expected behaviour

The user should be properly recognized by the server as being in the channel they were moved to.

Actual behaviour

User vanishes

Platform

Windows, Linux, Web, iOS, API, Self-hosting

17 comments

Sign in with Fluxer to comment and vote.
Comment by Rex
RexSystem 1 vote edited Merged from #434 2 replies
Status changed from Fixed to Awaiting confirmation
This may be fixed in the current version. If it still happens for you, please say so here.
Comment by Aidan
Aidan 1 vote
Sorry, didn't notice this when mine got merged in, but it is not fixed.
Comment by Excorio
This still occurs in my current flatpak version: Stable Desktop 2026.1105.11444 Web 2026.1107.204546 Linux 7.3.3-1-cachyos (x64) Electron 44.10.0 Chrome 152.0.7977.169 Node 24.22.1 Locale en-US Observed behavior:
  1. User joins a channel and connects to voice.
  2. An admin moves the user to another channel.
  3. Instead of being auto-connected in the new channel, the user is prompted to connect voice.
  4. Following this, the user is briefly presented with a reconnect prompt and gets placed back in the previous channel with voice connected.
My interpretation of the events: The sequence looks like the move is treated as a sudden disconnect, which appears to trigger an auto-reconnect feature (if it exists) that restores the last connected channel. If so, that feature makes sense for genuine sudden disconnects but should not fire on an intentional move. A moved use that was voice-connected in the previous channel and moved should be voice-connected in the destination channel they get moved to.
Comment by Aidan
Aidan 2 votes
Just started having a different experience. Now, any user that is dragged fails to connect to the channel and then is automatically put back in the channel they were dragged away from.
Comment by Zaostaly Jezevec
I can reproduce this on a self-hosted instance (API/gateway 2026.1114, desktop 2026.1114.60222 and web 2026.1114.52902) and found the root cause plus a working server-side workaround. Symptom PATCH /guilds/{g}/members/{u} with {"channel_id": ...} returns 200, the user sees "connecting" (orange), then an error, and after ~5 s the client puts them back in the original channel. Same on desktop and web. Console output (both clients):
[VoiceEngineV2AppConnectionHostAdapter] [Warn] Aborting LiveKit room creation after codec probing because attempt is stale {attemptId: 5}
  at handleVoiceServerUpdateAsync -> ... (VOICE_SERVER_UPDATE was received)
[VoiceEngineV2AppConnectionHostAdapter] [Warn] Voice server timeout {guildId, channelId: <target>}
[VoiceReconnectionManager] [Info] Attempting to reconnect to last voice channel {channelId: <original>}
The server side is fine: no errors in api/gateway/LiveKit, and the token request succeeds. Root cause: a race between two gateway events on the client After a move the gateway sends two events for the moved user almost at once:
  1. VOICE_STATE_UPDATE (new channel). The client handles it in MediaEngineFacade.ts (~L1889-1898): it calls disconnectForChannelMove('server') and, in .finally(), recoverConnectionExpectation(...). That starts a new connection attempt.
  2. VOICE_SERVER_UPDATE (token for the new channel). The handler in VoiceEngineV2AppConnectionHostAdapter.ts accepts it, then awaits getRoomVideoDecoderExclusions() and re-checks isLatestConnectionAttempt(attemptId) (~L519-526).
If (2) is processed while (1) is still in flight, the attempt id is bumped by recoverConnectionExpectation and the already running attempt for the token is discarded as stale. The server only sends the token once, so the client waits for a VOICE_SERVER_UPDATE that never comes again, times out after 5 s and reconnects to the old channel. With a fast, local LiveKit the token seems to arrive right after the state update, so this lost the race every time for us. (This is my reading of the client code, I did not trace the exact event order.) This also matches the "Connection succeeded but attempt is stale, disconnecting" log from other self-hosters in this thread. Workaround that works for us In fluxer_gateway/src/guild/voice/guild_voice_move.erl, delay the token dispatch a bit so VOICE_STATE_UPDATE is always handled first. The function already runs in a spawned process, so nothing is blocked:
-define(MOVE_TOKEN_DELAY_MS, 500).

send_captured_voice_server_updates(GuildId, ChannelId, SessionDataList, GuildPid) ->
    timer:sleep(?MOVE_TOKEN_DELAY_MS),
    lists:foreach(...)
With this single change the move worked on our instance (tested once, moving a user via the API; before the change it failed 3 out of 3 times on desktop and web). The fact that a pure delay fixes it supports the race explanation above. It is only a workaround: a proper fix would make the client tolerate either event order (e.g. not invalidate the attempt that is already waiting on the token for the same target channel), or have the server guarantee the order.
Comment by ternera
terneraMod 1 vote
Can reproduce CR - Canary Desktop 2026.606.140333, Web 2026.605.11941, macOS 26.5.1 (arm64), Electron 41.2.2, Chrome 146.0.7680.188, Node 24.14.1, Locale en-US
Comment by Comfy_Deer
Can reproduce CR - I had a similar report self-denied dupe #2991, except The User disappears while actually having joined the other VC channel. - Canary Desktop 2026.602.31138, Web 2026.605.11941, Linux 6.17.0-35-generic (x64), Electron 41.2.2, Chrome 146.0.7680.188, Node 24.14.1, Locale en-US
Comment by Fullest
FullestReviewer 1 vote
Can reproduce CR: Canary Desktop 2026.603.33821, Web 2026.605.11941, Linux 7.0.11-100.fc43.x86_64 (x64), Electron 41.2.2, Chrome 146.0.7680.188, Node 24.14.1, Locale en-US
Comment by @laurentm28
RexSystem 1 vote edited Merged from #838 originally by @laurentm28 on GitHub
This is closely related to #715 (voice channel move leaves the user stuck / LiveKit room transfer never fired server-side). That issue describes a disconnect on move; here the symptom is subtler: the UI moves the user, but the voice stream stays bound to the original channel's LiveKit room, so people in the original channel keep hearing them. Same root cause hypothesis: the destination channel's VOICE_SERVER_UPDATE / LiveKit token handoff never triggers. Worth fixing together. Repro (desktop, canary): User A in voice channel 1 → moderator with Move Members permission drags A to voice channel 2 → A disappears from channel 1 list, appears in channel 2, but A's audio is still heard in channel 1 and not in channel 2.
Comment by @laurentm28
RexSystem 1 vote Merged from #838 originally by @laurentm28 on GitHub
Additional data point from testing: the bug is not reproducible for every user. With the affected user it behaves like this: the first move (drag to another voice channel) works and the user visibly lands in the destination channel, but on a second consecutive move the user vanishes from the member list while their audio remains audible in the previous channel (and is not heard in the new one). So the failure seems tied to the state accumulated per-session (LiveKit room/connection) rather than to the move action itself — a fresh/healthy client survives the first transfer but a second transfer breaks it.
Comment by VII
VII 1 vote Merged from #2261
Can reproduce CR; also occurs without having to move someone, voice channels appear empty on mobile even in use. Version 2026.09.100, Beta 5, Channel Beta, Build 79, iOS 26.6.1 (arm64), Device iPhone17,4, Locale en-US, Push apns, Flutter 3.47.5, Dart 3.13.4
Comment by mesub
mesubMod 1 vote Merged from #2261
Can reproduce CR: Version 2026.09.100, Channel ALPHA, Build 2563, Android 15 (arm64), Device OnePlus CPH2465, Locale en-GB, Push unifiedpush, Flutter 3.47.5, Dart 3.13.4
Comment by penguin
Can reproduce this issue on a self-hosted Fluxer instance running the latest available version. Environment:
  • Self-hosted Fluxer
  • Canary Web 2026.1008.22644
  • Linux (x64), Firefox 157.0, Locale en-US
  • Second user: Chrome (web client)
  • Both users were using the web version, not the desktop app.
Observed behavior:
  • Moving another user to a different voice channel sometimes makes them disappear from the voice channel list entirely.
  • Moving them back sometimes works, but both users may then be unable to hear each other.
  • Repeated channel moves can also cause the moved user to disconnect.
  • Fully disconnecting and reconnecting both users temporarily resolves the issue.
Relevant log messages: Browser:
Connection throttled 2
Connection succeeded but attempt is stale, disconnecting
Server:
LiveKit participant_joined did not match gateway state
error: connection_not_found
Comment by probablyjassin
The way this reflects to me in communities is this: In my screenshot i was moved from the above VC to VC 2. The VC view is completely black, and the user doesn't show up in the VC members list. If another user joins this VC now, they are completely oblivious to the fact that another user is there. But the moved user can't hear them, because if the incoming user is unmuted, a loud set of garbled noises permanently comes out the moved users audio outputs, as if it was coming from the incoming users microphone. Canary Web 2026.1010.212806, Linux (x64), Firefox 157.0, Locale en-US
Comment by Aidan
Aidan 1 vote Merged from #2261
Have also reproduced, still a problem. Reproduced on main server using iOS, Linux, and Windows. Creating new full ticket given this seems to be a server side issue.