Merge pull request #183703 from anantshankar17/master

PRMerger8 · web-flow · commit 7f2f2364503b · 2022-03-16T06:41:23.000-07:00
Max value of InstanceCloseDelay
diff --git a/articles/service-fabric/service-fabric-application-upgrade-advanced.md b/articles/service-fabric/service-fabric-application-upgrade-advanced.md
@@ -91,6 +91,7 @@ The overridden delay duration only applies to the invoked upgrade instance and d
 > * The settings to drain requests will not be able to prevent the Azure Load balancer from sending new requests to the endpoints which are undergoing drain.
 > * A complaint based resolution mechanism will not result in graceful draining of requests, as it triggers a service resolution after a failure. As described earlier, this should instead be enhanced to subscribe to the endpoint change notifications using [ServiceNotificationFilterDescription](/dotnet/api/system.fabric.description.servicenotificationfilterdescription).
 > * The settings are not honored when the upgrade is an impactless one i.e when the replicas will not be brought down during the upgrade.
+> * The max value of InstanceCloseDelayDuration that can be configured in the service description or the InstanceCloseDelayDurationSec in the upgrade description can't be greater than cluster config FailoverManager.MaxInstanceCloseDelayDurationInSeconds, which defaults to 1800 seconds. To update the max value, the cluster level config should to be updated. This configuration is only available in the runtime version 9.0 or later.
 >
 >
 

Original file line number	Diff line number	Diff line change
`@@ -91,6 +91,7 @@ The overridden delay duration only applies to the invoked upgrade instance and d`
`91`	`91`	`> * The settings to drain requests will not be able to prevent the Azure Load balancer from sending new requests to the endpoints which are undergoing drain.`
`92`	`92`	`> * A complaint based resolution mechanism will not result in graceful draining of requests, as it triggers a service resolution after a failure. As described earlier, this should instead be enhanced to subscribe to the endpoint change notifications using [ServiceNotificationFilterDescription](/dotnet/api/system.fabric.description.servicenotificationfilterdescription).`
`93`	`93`	`> * The settings are not honored when the upgrade is an impactless one i.e when the replicas will not be brought down during the upgrade.`
	`94`	`+> * The max value of InstanceCloseDelayDuration that can be configured in the service description or the InstanceCloseDelayDurationSec in the upgrade description can't be greater than cluster config FailoverManager.MaxInstanceCloseDelayDurationInSeconds, which defaults to 1800 seconds. To update the max value, the cluster level config should to be updated. This configuration is only available in the runtime version 9.0 or later.`
`94`	`95`	`>`
`95`	`96`	`>`
`96`	`97`