Load Balancer Dedicated Cluster Related Issues

I have purchased a dedicated cluster, why can't I see it when I create a BLB instance?

1. First, verify that the dedicated cluster is in available status. A newly purchased dedicated cluster will automatically transition to available status after the establishment of clusters (visible on the cluster list page). If not listed, the cluster is still under establishment, please wait.
2. Secondly, if the cluster is in available status, verify that the switch on the Cluster Details page is enabled. If the switch is disabled, the cluster becomes unselectable which prohibits the creation of BLB instances.
3. To select this cluster, redirect to its details page and enable the switch

For BLB instances created on dedicated clusters, how to choose between active-standby mode and active-active mode?

1. For dedicated clusters spanning two different data centers, instances created on the cluster support active-standby mode/active-active mode. The active-standby mode is supported by default.
2. Both active-standby and active-active modes relate to client access patterns for BLB data flows. In active-standby mode, regardless of which cloud data center initiates the access, requests will be send to the primary data center of the dedicated cluster of the corresponding instance by default. The traffic will be redirected to the standby center only when the primary data center fails. The active-active mode follows proximity access. The traffic will be forwarded to the nearest data center of the dedicated cluster from the client’s origin based on the shortest physical line principles.
3. If the BLB's real servers are located within a single cloud data center and you require real link affinity for BLB access, it is recommended to use active-standby mode and assign the primary data center to match or be physically closer to the real servers. If there are no restrictions on data center locations for clients accessing the BLB instance and forward link affinity is preferred, then active-active mode can be selected.
4. From the product respect, it is recommended to select active-standby mode for clearer data flow, easier assessment of load across data centers of the dedicated cluster, and more precise localization of faulty data flow link.

When BLB instances migrate between dedicated clusters, shared clusters or different dedicated clusters, will data flow be lossy? How long will the expected lossy duration be?

1. The lossiness of both long and short link data flows will affect accesses to the BLB instance's Layer 4 listener. The client must actively reconnect to recover. The lossiness duration depends on the listener's health check parameters (the duration equals to health check interval × healthy threshold).
2. To minimize the lossiness impact of migration, it is recommended to perform the migration during off-peak hours or after disabling the service. Moreover, these two health check parameters of the listener should be set to the minimum before migration, and then restored to the original after migration.

Will data flow be lossy during switching of BLB instances in a dedicated cluster from active-standby data centers to active-standby or active-active mode? How long will the expected lossy duration be?

For Layer 4 listener access on the BLB instance, new links will not experience packet loss. However, existing long and short links may encounter packet loss during switches between active-standby data centers or probabilistic packet loss during switches between active-standby/active-active data centers. Clients can recover by re-establishing connections. The lossiness duration depends on the retry time configured on the client side.