Peter Steinberger 5232aba003 fix(windows): recover Gateway restarts from SQLite sharing errors (#159347) ...

* fix(windows): recover Gateway restarts from SQLite sharing errors

Ask the verified Gateway owner to drain before Task Scheduler termination, wait for task completion, and retry transient Windows lease reads. Preserve restart activation with a warning when post-stop inspection remains unavailable and the port is free. Thanks @eddiekk for the failure traces.

* test(windows): tie stop fixtures to process termination

Route Scheduled Task probes independently of CIM observation, retain listeners until their owned process is terminated, and revoke continuation authority at the intended post-End phase. Preserve owner termination and foreign-process protection assertions.

* fix(windows): bind restart cleanup and wait for task settlement

Keep cleanup bound to captured process identities, preserve later replacements,
and wait for the Scheduled Task to settle before activation. Record recovery
when transient SQLite owner inspection remains locked after confirmed exit.

* fix(windows): preserve captured children and report refused stops

* fix(windows): recover restart after unavailable task inspection

* fix(android): regenerate gateway stop protocol method

2026-09-28 01:18:31 +00:00

..

discovery.md

…

query.md

fix(daemon): inspect registered Gateway services accurately (#151608)

2026-09-27 02:55:51 -07:00

restart-and-supervision.md

fix(doctor): report remaining systemd units after cleanup (#158034)

2026-09-25 15:42:19 +00:00

running.md

fix: keep gateway startup retryable after failed reads (#159598)

2026-09-27 11:08:30 -07:00

service.md

fix(windows): recover Gateway restarts from SQLite sharing errors (#159347)

2026-09-28 01:18:31 +00:00