microsoft/aspire · error · InvalidOperationException

Cannot reconnect: no previous connection.

Error message

Cannot reconnect: no previous connection.

What it means

AppHostCliBackchannel can reconnect to a new AppHost process only if it remembers the Unix domain socket path from a previous connection. When ReconnectInternalAsync runs (triggered from OnDisconnected) before any connection was ever established, the socket path is null, so the CLI throws this InvalidOperationException rather than attempting a reconnect to an unknown target.

Solutions

1. Ensure the initial connection is established (and thus _socketPath is set) before disconnect/reconnect handling is subscribed to or can fire.
2. Guard reconnect logic: check whether the backchannel has ever connected before attempting reconnection.
3. If the AppHost died before connecting, treat it as a fresh start (start/restart the AppHost and connect anew) rather than a reconnect.

Example fix

// before
await backchannel.ReconnectInternalAsync();

// after
if (backchannel.HasEverConnected)
{
    await backchannel.ReconnectInternalAsync();
}
else
{
    await backchannel.ConnectAsync(cancellationToken);
}

if (backchannel is not { HasEverConnected: true }) { /* treat as fresh connect, not reconnect */ }

bool canReconnect = backchannel is { HasEverConnected: true };

try { await backchannel.ReconnectInternalAsync(); }
catch (InvalidOperationException ex) when (ex.Message.Contains("no previous connection"))
{ await backchannel.ConnectAsync(ct); }

When it happens

Trigger: A disconnect notification (OnDisconnected) fires on the backchannel before any successful ConnectAsync completed, so _socketPath was never populated. Calling ReconnectInternalAsync directly on a freshly constructed AppHostCliBackchannel also throws this.

Common situations: The AppHost process dies or drops the socket very early in CLI startup, racing with the initial connect; a consumer calls reconnect logic before the first connection; a race between disconnect events and connection establishment.

Understand the failure class

Background: "Invalid state transition" errors: "status must be X, actually Y", "already rejected/charging/uninstalled", "cannot ... while running" — what they mean when a library rejects your call — this error's family across 31 libraries.

Related errors

• Already connected to
• Already connected to AppHost backchannel.
• AppHost is incompatible with the CLI. The AppHost must be…
• Not connected to auxiliary backchannel.
• Resource update buffering was not enabled for this watcher.

AI-assisted analysis of microsoft/aspire@25830f84bd (2026-09-16). Data as JSON: /api/errors/b6adaab0e08135b0. Report an issue: GitHub.

Appendix: source

Thrown at src/Aspire.Cli/Backchannel/AppHostCliBackchannel.cs:448

        {
            _disconnectTaskCompletionSource.TrySetResult();
        }
    }

    private void ResetForReconnection()
    {
        lock (_lock)
        {
            logger.LogDebug("Resetting backchannel for reconnection");
            _rpcTaskCompletionSource = new TaskCompletionSource<JsonRpc>();
        }
    }

    private async Task ReconnectInternalAsync()
    {
        if (_socketPath is null)
        {
            throw new InvalidOperationException("Cannot reconnect: no previous connection.");
        }

        ResetForReconnection();

        // Wait for the new socket to appear (the new DistributedApplication needs to start)
        var startTime = DateTime.UtcNow;
        var maxWait = TimeSpan.FromSeconds(30);

        var retryCount = 0;
        while (!_cancellationToken.IsCancellationRequested)
        {
            try
            {
                await ConnectAsync(_socketPath, _autoReconnect, retryCount, _cancellationToken).ConfigureAwait(false);
                logger.LogInformation("Successfully reconnected to backchannel");
                return;
            }
            catch (SocketException) when (DateTime.UtcNow - startTime < maxWait)

View on GitHub (pinned to 25830f84bd)