Repository navigation
Conversation
A recreated global service (nginx-proxy, redis, cron, newrelic-daemon) got 60 s to become ready before the update rolled back. Slow or heavily loaded servers can take longer, so raise the default to 180 s. The database keeps its 300 s, and `global-service-ready-timeout` in config.yml still overrides both.
|
No actionable comments were generated in the recent review. Walkthrough
Change: Bug fix Merge Risk:
|
Summary
GlobalContainers::READY_TIMEOUT).nginx -t, a response on port 80, then a 2 s settle). The update then rolls back to the previous proxy andee cli updatefails, although nothing is wrong except speed.global-service-ready-timeoutinconfig.ymlstill overrides both. A healthy proxy is still accepted as soon as it is ready (2–3 s in the v4.14.1 upgrade tests), so normal updates don't get slower. Only a proxy that never comes up now takes up to 180 s before the rollback, instead of 60 s.Planned for v4.14.2.
Testing
php -lpasses. The only change is the constant; the timeout logic and its config override are unchanged from v4.14.0.Summary by CodeRabbit
global-service-ready-timeoutsetting can still override this default. The database readiness timeout remains unchanged at five minutes.