fix(mcp): widen discovery lock wait to cover real discovery durations

The bounded lock wait was 10 x 0.2s = 2s, but the concurrent-discovery
scenario this lock exists for (#62771: hermes serve + gateway spawning
every MCP server twice) reports 40-60s discovery rounds. A 2s budget
guarantees the loser times out and runs unguarded exactly when the guard
matters. Widen to 240 x 0.5s = 120s; fail-soft unguarded fallback and
test overrides unchanged.
This commit is contained in:
Teknium 2026-07-28 12:02:10 -07:00
parent 362fc21577
commit 8151670712

View file

@ -4106,8 +4106,8 @@ _LOCK_UNAVAILABLE: Any = object() # sentinel: locking broken/unavailable
_MCP_DISCOVERY_LOCK_PATH: Optional[str] = None # resolved lazily
# Retry constants for the bounded wait when another process holds the lock.
_MCP_DISCOVERY_LOCK_MAX_RETRIES: int = 10
_MCP_DISCOVERY_LOCK_RETRY_DELAY_S: float = 0.2
_MCP_DISCOVERY_LOCK_MAX_RETRIES: int = 240
_MCP_DISCOVERY_LOCK_RETRY_DELAY_S: float = 0.5
class _LockCookie: