You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Servers backup - if an upstream is down and backup server is configured, it should get the traffic. When the upstream backs to normal operation, the backup should not receive traffic.
At least following functionality must be covered by the functional test:
HTTP scheduler working over at least 100 server groups emulating 100 virtual hosts;
servers failovering - servers must get load when they return to normal operation while the service must stay working during failovering. This point depends on [HTTP] Don't pipeline non-idempotent method requests #419 and Large number of failures in ab benchmark tests #383: currently the system returns errors if upstream connection is reset. After fix of the issues there should not be such errors.
Servers backup - if an upstream is down and backup server is configured, it should get the traffic. When the upstream backs to normal operation, the backup should not receive traffic.
All the schedulers, including Sticky sessions scheduler #593 and Advanced Load Balancing #565 must be involved in the test.
Some stress must be put onto the configurations to verify that there is no crashes. So some traffic generator like
wrk
should be used.The text was updated successfully, but these errors were encountered: