Signet: node connection returns TLS alert 225 (SchedulerError) after scheduler schedule succeeds #750

Context: Greenlight backend for Alby Hub (https://github.com/proxy/gist.github.com/welliv/cf1f9b22a3b9047e5a31f6dc4cc2de24). Regtest passes; signet blocks live testing.

Client: glcli 0.1.3 and gl-client built from main @ 15b22cf, with a developer certificate.

Nodes

• Signet A: 0232c2db24b4eb134eacacd0a5e1ff16a5ef7848958ed36114b374d6728cb7c5d3
  (gl1qgev9keykn43xn4v4ng2tc0lz6j777zgjk8dxcg5kd6dvu5vklzax75ccyz.nodes.gl.blckstrm.com)
• Signet, freshly registered: 03634f9c7ba66adee0ea8f481bd3ac64e244bc91b24ee37d5d4b5ea92e3050d701

Repro

1. glcli -n signet scheduler register (fresh node only), then glcli -n signet signer run
2. glcli -n signet scheduler schedule → returns grpc_uri
3. glcli -n signet node getinfo → alert 225

Timeline (2026-10-11, UTC)

Evidence excerpts

[2026-10-11T06:10:18Z DEBUG tonic::transport::service::reconnect] reconnect::poll_ready: hyper::Error(Connect, Custom { kind: InvalidData, error: AlertReceived(Unknown(225)) })
error trying to connect: received fatal alert: Unknown(225)

>>> TLS 1.3, Handshake [length 0615], ClientHello
<<< TLS 1.3, Alert [length 0002], fatal ???
...SSL alert number 225

What we ruled out

• Clock, TLS version, CA, missing client cert, SNI and --network flag. The same machine, binary and cert family connect fine to a testnet3 node.
• DNS. A direct connection to 34.149.188.3 gives the same alert.

Your node-domain docs list 225 as SchedulerError ("the scheduler was unable to locate or schedule the node"), a server-side error to report to you. Hence this issue.

Question

Is signet scheduling currently available for developer-certificate nodes? If it is, the node ids and times above should match your scheduler logs.