connect: cannot attach to an existing healthy hub (EADDRINUSE)
Summary
A restarted cline connect slack refuses to attach to an already-running, healthy hub. Instead of connecting to the hub listening on port 25463, it tries to spawn its own hub-daemon, which immediately dies with EADDRINUSE because the healthy hub already owns the port.
Impact
The connector cannot start at all while a working hub is running. The only recovery found was to kill the healthy hub so the new connect invocation could spawn a hub of its own — which needlessly tears down every other session that hub was serving.
Steps to reproduce
- Have a healthy hub running and serving sessions on port 25463.
- Restart
cline connect slack(or start a second one). - The new process does not attach to the existing hub; it attempts to spawn
hub-daemon. - That spawn fails with
EADDRINUSEon 25463 and the connector exits.
Expected
connect should detect the healthy hub already listening on the port and attach to it, rather than unconditionally spawning its own daemon and failing. A port that is in use by a working hub is the normal case for a restart, not an error condition.
Environment
- Observed on 3.0.47 and 3.0.46-nightly
- Dates observed: 2026-07-29 and 2026-07-30
- Hub port: 25463
Source: cline/cline