Suboptimal health monitoring

Topics

• Host suboptimal health

• Network suboptimal heath

Host suboptimal health

Host suboptimal health means that the host has no obvious faults during operation, but its

performance metrics or operating status are abnormal and pose potential risks. A host in suboptimal health status might have abnormal resource usage, slower response, or hardware or software risks. If you do not process these issues in time, they might cause system faults or service interruptions. After you configure suboptimal health check on the HA page of the cluster, the system automatically identifies and displays hosts with suboptimal health (at risk of becoming unresponsive or already unresponsive). When the system detects a suboptimal host during HA, it downgrades that host. VMs do not preferentially run on suboptimal hosts. After you manually fix a system with suboptimal health, you can release the suboptimal host and restore its preference.

Features

• Suboptimal host list

• Suboptimal health issues

Suboptimal host list

Use this feature to display hosts in a suboptimal health status on the page. You can quickly search for

a destination host by host name or IP address. You can manually remove generated hosts from the suboptimal host list.

1. On the top navigation bar, click HA Service.

2. From the navigation pane, select Suboptimal Health Monitoring > Host Suboptimal Health.

Suboptimal health issues

On this page, you can view the specific suboptimal health issues of each host. You can also quickly

search suboptimal hosts by response action, host name, or host IP.

1. On the top navigation bar, click HA Service.

2. From the navigation pane, select Suboptimal Health Monitoring > Host Suboptimal Health.

3. Click the Suboptimal Health Issues tab.

Network suboptimal heath

In RBAC mode, only the system administrator can configure this feature.

The network performance of the management platform will be in critical state after it has been operating for a long time. As a result, the management platform might fail or its performance might degrade. Suboptimal network health might seriously affect service stability and user experience. You must detect network issues and resolve them immediately. The system can monitor and analyze performance metrics of vSwitch NICs, including the latency, packet loss rate, and abnormal flappings, optical power anomalies, and CRC error rate of physical links. It promptly generates alarms about various suboptimal health conditions. The system checks only the management, storage, and service networks, and quarantines faulty network links when necessary to prevent the network from being in suboptimal health state for a long time, ensuring stable and efficient operation of the entire network.

Configure a suboptimal health check policy

Perform this task to flexibly set check metrics, and send and process alarms about suboptimal health issues for vSwitch NICs.

1. On the top navigation bar, click HA Service.

2. From the navigation pane, select Suboptimal Health Monitoring > Network Suboptimal Health.

3. Click Configure.

4. In the dialog box that opens, select a cluster and enable suboptimal health check.

   ○ Cluster: Select the cluster for which the network suboptimal health check policy is configured. If you select All, the policy applies to all clusters matching the current version. You can select only the clusters that match the current version.

   ○ Suboptimal Health Check: Select whether to enable suboptimal health check. With this option enabled for a cluster, the system periodically checks the network health status of hosts in that cluster. When a vSwitch physical link has high latency, packet loss, jitter, or abnormal optical power, the system generates alarms and quarantines the abnormal link. If you do not enable this option, the system does not perform the suboptimal health check.

5. Configure network suboptimal health check policies by network type.

   ○ Management: Check suboptimal health status of the management network. Make sure the names of vSwitches between hosts in the cluster are consistent.

   ▪ Hardware Check: Turn on this option to perform hardware check for host NICs on the management platform.

   ▪ Hardware Check Sensitivity: Set the threshold for detecting virtual NIC exceptions within a detection interval and the number of consecutive faulty intervals. A higher sensitivity indicates a stricter check and an increased possibility of false alarms.

▪ Packet Sending Check: Enable this option to send probe packets to host NICs for checking the link packet loss and latency.

▪ Packet Loss Sensitivity: Select the sensitivity for packet loss on the host NIC.

▪ Latency Sensitivity: Select the sensitivity for communication latency on the host NIC.

○ Storage: Check suboptimal health status of the storage network. The configuration items are the same as those on the management network, including hardware check, hardware check sensitivity, packet sending check, packet loss sensitivity, and latency sensitivity. Make sure the names of vSwitches between hosts in the cluster are consistent. The system only triggers an alarm notification and does not perform a quarantine when the following conditions are met:

• RDMA is enabled and link aggregation is configured for the storage network.

• A link aggregation member has an exception.

○ Service: Check suboptimal health status of the service network. Service networks do not support latency and packet loss checks. Only configuration related to hardware check is supported.

▪ Hardware Check: Turn on this option to perform hardware check for host NICs on the management platform.

▪ Hardware Check Sensitivity: Set the threshold for detecting virtual NIC exceptions within a detection interval and the number of consecutive faulty intervals. A higher sensitivity indicates a stricter check and an increased possibility of false alarms.

○ Storage Server: Check communication quality between the cluster host and the storage server.

▪ Packet Loss Sensitivity: Select the sensitivity for packet loss in communication within the check interval.

▪ Latency Sensitivity: Select the sensitivity for latency in communication within the check interval.

6. Click OK to save and apply the network suboptimal health check policies.

Clear quarantine for a NIC

If a quarantined NIC has no fault, you can clear quarantine for it and add it to the aggregation group again.

1. On the top navigation bar, click HA Service.

2. From the navigation pane, select Suboptimal Health Monitoring > Network Suboptimal Health.

3. In the Exception Info list, click Resume in the Actions column for the target NIC.

Turn on the NIC LED

Use this feature to have the NIC LED blink regularly so you can quickly identify and locate the NIC. On the same host, you can turn on only one NIC LED at a time.

1. On the top navigation bar, click HA Service.

2. From the navigation pane, select Suboptimal Health Monitoring > Network Suboptimal Health.

3. In the Exception Info list, click Turn On NIC LED in the Actions column for the target NIC.

4. In the dialog box that opens, click OK.

Turn off the NIC LED

Use this feature to turn off a NIC LED.

1. On the top navigation bar, click HA Service.

2. From the navigation pane, select Suboptimal Health Monitoring > Network Suboptimal Health.

3. After replacing a NIC, in the Exception Info list, click Turn Off NIC LED in the Actions column for the target NIC.

View network suboptimal heath check summary

Use this feature to view information about used NICs on the management platform, a summary of network suboptimal health issues, and alarm events triggered by network suboptimal health.

1. On the top navigation bar, click HA Service.

2. From the navigation pane, select Suboptimal Health Monitoring > Network Suboptimal Health.

3. On the page that opens, you can view suboptimal health issue summary, network exceptions, and vSwitch network information.

   ○ The Exceptions area displays the most recent five alarms by default. To view all alarms, click More.

   ○ On the Exceptions page, you can set filter criteria, and click Filter to filter alarms.

   ○ To delete alarms that have been cleared, select the alarms, and then click Delete.

   ○ To bulk confirm alarms, select the alarms, and then click Confirm.

   ○ To confirm a specific alarm, click Confirm in the Actions column for that alarm.

4. To access the real-time alarms page, click Alarm.