stablyai/orca · warning

SSH provider wait aborted for target "${targetId}".

Error message

SSH provider wait aborted for target "${targetId}".

What it means

Thrown by waitForRuntimeSshProviders when the caller-supplied AbortSignal becomes aborted during the up-to-10s polling loop that waits for both the SSH git provider and SSH filesystem provider to register for a freshly connected target. The connect attempt in connectRuntimeOwnedSshTarget is being cancelled mid-flight, so the runtime-owned SSH target is removed (cleanup) and the abort error is rethrown. It is a cooperative cancellation signal, not a transport failure.

Source

Thrown at src/main/ephemeral-vm-runtime-ssh.ts:63

export async function disconnectRuntimeOwnedSshTarget(targetId: string | undefined): Promise<void> {
  if (!targetId) {
    return
  }
  await disconnectRegisteredSshTarget(targetId)
}

export async function removeRuntimeOwnedSshTarget(targetId: string | undefined): Promise<void> {
  if (!targetId) {
    return
  }
  await removeRegisteredSshTarget(targetId)
}

async function waitForRuntimeSshProviders(targetId: string, signal?: AbortSignal): Promise<void> {
  const startedAt = Date.now()
  while (Date.now() - startedAt < SSH_PROVIDER_READY_TIMEOUT_MS) {
    if (signal?.aborted) {
      throw new Error(`SSH provider wait aborted for target "${targetId}".`)
    }
    if (getSshGitProvider(targetId) && getSshFilesystemProvider(targetId)) {
      return
    }
    await new Promise((resolve) => setTimeout(resolve, SSH_PROVIDER_READY_INTERVAL_MS))
  }
  throw new Error(`SSH relay providers were not ready for target "${targetId}".`)
}

View on GitHub (pinned to 1136503c6a)

Solutions

1. If the abort is expected (user cancelled), surface a 'Launch cancelled' state in the UI and stop retrying — do not treat this as a transient transport error.
2. If the abort is coming from an overly tight caller timeout, lengthen or remove the caller's AbortSignal so the 10s SSH_PROVIDER_READY_TIMEOUT_MS budget can do its job.
3. If providers are genuinely slow to register and you need more time, that is a different error (1041) — do not paper over it by aborting; investigate relay registration latency instead.
4. Ensure only one connect attempt owns a given runtimeId at a time so a winning attempt does not abort a still-valid in-flight one.

Example fix

// before
const ctrl = new AbortController()
setTimeout(() => ctrl.abort(), 2_000) // too tight for relay bring-up
await connectRuntimeOwnedSshTarget({ runtimeId, connection, signal: ctrl.signal })

// after: let the provider-wait budget run; abort only on real user cancel
await connectRuntimeOwnedSshTarget({ runtimeId, connection, signal: userCancelSignal })

import { getSshGitProvider, getSshFilesystemProvider } from './providers/ssh-dispatch'

// Before connect, there is nothing to validate (target not created yet).
// Instead, scope the abort to real user cancellation only.
const userCancel = new AbortController()
document.getElementById('cancel-launch').onclick = () => userCancel.abort()
// Do NOT add a short setTimeout(...).abort() here — the 10s provider-wait budget is intentional.

function isSshProviderWaitAborted(error: unknown): boolean {
  return error instanceof Error && error.message.startsWith('SSH provider wait aborted for target ')
}

try {
  await connectRuntimeOwnedSshTarget({ runtimeId, connection, signal: userCancel.signal })
} catch (error) {
  if (isSshProviderWaitAborted(error)) {
    setLaunchState('cancelled')  // user-cancelled, not a transport fault
    return
  }
  throw error
}

When it happens

Trigger: connectRuntimeOwnedSshTarget is called with an AbortSignal (e.g. an ephemeral VM recipe is cancelled, the user closes the launching dialog, or a higher-level operation races and wins). During the <=10s window where getSshGitProvider(targetId) and getSshFilesystemProvider(targetId) are still undefined, signal.aborted flips true and the next loop iteration throws at ephemeral-vm-runtime-ssh.ts:63.

Common situations: User cancels an ephemeral VM / SSH recipe launch while the relay is still bringing up providers; a UI timeout on the launch surface aborts the connect; concurrent connect attempts to the same runtime where one wins and the other is aborted; tests that pass a short AbortSignal and do not await provider registration.

Related errors

• SSH relay providers were not ready for target "${targetId}".
• SSH target did not connect: ${state.status}
• Clone aborted
• ${response.error.message}
• Couldn't verify the SSH connection. Reconnect the host and t

AI-assisted analysis of stablyai/orca@1136503c6a (2026-08-12). Data as JSON: /api/errors/5611024539c260a4. Report an issue: GitHub.