Skip to main content

Check system health

Perform this task to check the system for operating and performance issues and any other risks that

might negatively impact VM services and rate the system health state. System health check provides information about the system, compute resources, storage resources, network resources, and HA service state. On detection of any exceptions, the system automatically provides recommended actions to help you maintain or restore system health. The detection results can be printed and exported.

Start a health check

  1. Select Hotkey > Health Check.

  2. Select health check options, and then click Start. You can click Select All to select all options. After completing the health check, the system displays the number of check options and numbers of failures and alarms, if any. To switch between the check item list, normal item list, and abnormal item list, click All Items, Normal, or Abnormal.

  3. To cancel a health check, click Stop.

  4. To perform a new health check, click Recheck, reselect options, and then click Start.

Configure health check settings

  1. Select Hotkey > Health Check.

  2. Click Check Settings.

  3. Configure the parameters as described in "Parameters."

  4. Click OK.

Edit health check settings

  1. Select Hotkey > Health Check.

  2. Click Check History.

  3. Click the target health check task, and then click Edit.

  4. Edit the settings as required, and then click OK.

View health check history

  1. Select Hotkey > Health Check.

  2. Click Check History.

  3. Click the target history health check task.

Delete a history health check task

  1. Select Hotkey > Health Check.

  2. Click Check History.

  3. Click the target history health check task, and then click Delete.

  4. Click OK.

Print a history health check task

  1. Select Hotkey > Health Check.

  2. Click Check History.

  3. Click the target history health check task, and then click Print.

Export a history health check task to a PDF file

  1. Select Hotkey > Health Check.

  2. Click Check History.

  3. Click the target history health check task, and then click Export as PDF.

Export a history health check task to an Excel file

  1. Select Hotkey > Health Check.

  2. Click Check History.

  3. Click the target history health check task, and then click Export as Excel.

Parameters

System: Checks the running status of the resources and basic services of the management platform.

• CPU Overcommitment: Checks the ratio of assigned vCPUs to physical CPU cores (including those cores created by using hyperthreading) on each compute node on the management platform. If that ratio exceeds 100%, unstable server operation and unexpected errors will occur. In a virtualized environment, high CPU overcommitment can cause slow response or even abnormality of the services provided by the VMs running on a server. As a best practice, limit the CPU overcommitment ratio to be below 200%.

• Memory Overcommitment: Checks the ratio of assigned virtual memory to physical memory on each compute node on the management platform. If that ratio exceeds 100%, unstable server operation and unexpected errors will occur. In a virtualized environment, high memory overcommitment can cause slow response or even abnormality of the services provided by the VMs running on a server. As a best practice, limit the memory overcommitment ratio to be below 80%.

• Storage Overcommitment: Checks the ratio of assigned virtual disk capacity to physical disk capacity on each shared storage volume on the management platform. If that ratio exceeds 80% on a shared storage volume, VMs that use this storage volume cannot perform disk I/O read or write operations, and the services provided by the VMs will be interrupted. As a best practice, limit the storage overcommitment ratio to be below 80%.

• System Alarm State: Checks for unrecovered and unacknowledged system alarms on the management platform.

• System Partition Usage: Checks the usage of system partitions, including root, log, and local file partitions on CVM and CVK.

○ Normal: The partition usage is less than 80%.

○ Warning: The partition usage is equal to or greater than 80% but less than 90%.

○ Failure: The partition usage is equal to or greater than 90%.

• NTP Settings: Checks whether NTP is configured for hosts. NTP enables all hosts to synchronize time with the CVM to ensure that the basic functions such as HA and system log timestamps can operate correctly.

• System Data Backup Settings: Checks whether scheduled data backup is configured for CVM. Scheduled data backup enables fast data recovery when CVM fails.

• Software Version Consistency: Checks software consistency between CVM and CVK.

• VMPtools State: Checks whether all VMs have VMPtools installed and the VMPtools is operating correctly. VMPtools provides the performance monitoring, application monitoring, failure monitoring, and clock synchronization functions. If VMPtools is not installed or is operating incorrectly, performance monitoring data correctness, application software reliability, VM OS blue screen HA, and VM OS internal clock synchronization will be affected.

• VM Anti-Virus Settings: Checks whether anti-virus is enabled for all VMs. VM disks and network I/O traffic can be protected in real time when anti-virus is enabled. The anti-virus function requires installation of VMPtools and the VMPtools must be operating correctly.

• Alarm Output Settings: Checks for system alarm settings, including third-party, mail, and SMS alarm settings. System alarms allow the administrator to monitor cluster, host, and VM usage, network and storage connectivity, security, and system failure and locate problems.

• License Registration State: Checks management platform license registration. By default, all licenses are trial licenses with a 180-day trial period. When the trial period expires, you cannot log in to CVM, but VM services can still run correctly.

• CVM Failover Configuration: Checks whether stateful failover is configured in CVM and whether stateful failover is operating correctly.

• Snapshot Configuration Check: Checks for VMs that use internal snapshots. You can take only internal snapshots for a VM that already has internal snapshots, and the VM will be suspended during internal snapshot creation, which might interrupt services. If files are widely dispersed on the VM disks, creation and deletion of internal snapshots will be very slow. It checks for snapshot policies with a retained snapshot limit larger than 3 and checks the allowed maximum number of snapshots. The storage pool cannot assure VM service creation if too many external snapshots are created for VMs.

Compute

• Compute Node State: Checks the operating status of the compute node hardware. A compute node can be in Normal, Maintenance, Abnormal, or Quarantined state. Normal indicates that the compute node is operating correctly and can be connected to and managed by CVM. Maintenance indicates that the compute node is in manual hardware maintenance state, such as motherboard or memory module replacement. Abnormal indicates that the compute node has failed, or a connectivity issue has occurred between the compute node and CVM. Quarantine indicates that the compute node has been quarantined. If a compute node is in this state, you cannot create, migrate in, or clone VMs, but you can shut down or delete VMs.

• CPU Hardware State: Checks the usage, temperature, and frequency of physical CPUs on the compute node.

○ Normal: The physical CPU usage is less than 60%.

○ Warning: The physical CPU usage is equal to or greater than 60% but less than 80%.

○ Failure: The physical CPU usage is equal to or greater than 80%.

• Memory Hardware State: Checks the installation state, usage, and swap partition usage of physical memory hardware on the compute node. If the memory hardware is not in position, for example, the memory module is loose, the host might get stuck, go down, or reboot. If the memory usage or swap partition usage is high, the system performance of VMs might be affected.

• VM State: Checks the state of all VMs on the management platform, including Normal, Shutdown, and Abnormal.

• DRS Configuration: Checks whether compute DRS is enabled for a cluster. Compute DRS balances CPU and memory resources on each compute node in a cluster. If the CPU or memory usage of a compute node is higher than the DRS threshold for a specific period of time (5 minutes by default), the compute resources will be automatically balanced in the cluster based on the DRS algorithm.

Storage

• Physical Disk State：Checks the state of physical disks, physical disk size, and physical disk temperature.

• Logical Disk State：Checks the state of logical disks, Logical disk size, Logical disk RAID level and Physical disk quantity.

• Storage Usage State: Checks the usage of shared storage, including distributed storage, SAN, and NAS. Insufficient shared storage space might cause VM service abnormality.

○ Normal: The shared storage usage is less than 60%.

○ Warning: The shared storage usage is equal to or greater than 60% but less than 80%.

○ Failure: The shared storage usage is equal to or greater than 80%.

• RAID Controller State：Checks whether the RAID controller is operating correctly, and RAID controller battery state. If the RAID controller or RAID controller battery state is abnormal, identify whether the RAID controller or physical disk is faulty. If it is, contact the server vendor to replace the RAID controller or physical disk. Hosts running a lower software version (for example, a version earlier than E0920) do not support the RAID controller status check item. Therefore, the health check result for such hosts is empty.

Network

• Physical NIC State: Checks the state of each physical NIC on a compute node, including whether the NIC driver is operating correctly, whether the NIC is in position, and whether NIC negotiation is operating correctly. Physical NIC driver failure or absent NICs will affect access to cluster management and VM services. Physical NIC rate and duplex negotiation failure will affect VM traffic forwarding efficiency.

• Virtual Interface State: Checks vPort state on virtual switches, including Normal, Alarm, and Failure.

• Network Service Process: Checks network service and database process state and CPU usage. For the ovsdb-server process, the system provides only state information. Its CPU usage is not checked.

• Network Packet Loss Statistics: Checks the packet loss and error packet rates within 10 seconds for vPorts and physical ports. If the packet loss rate or error packet rate is greater than 0.1%, VM service access performance will decrease.

• Host Route: Checks for multiple default routes or duplicated routes on the hosts on a compute node. Multiple default routes or duplicated routes might cause VM communication failure.

• Aggregate State: Checks the link aggregation mode, including Dynamic, Static, and Not Configured.

• Network Connectivity:

○ Management network connectivity: Identifies whether hosts can connect to the management network correctly.

○ Non-management network connectivity: Identifies whether hosts can connect to a non-management network correctly. The check result displays the vSwitch name and subnet name. If no subnet exists, the vSwitches with the same name are displayed by default.

• Subnet Mask/Prefix Consistency Check: Checks subnet mask/prefix consistency among vSwitches with the same name in a cluster. If their subnet masks/prefixes are not the same, they cannot communicate with each other.

• MTU Consistency: Checks MTU consistency among vSwitches with the same name in a cluster and between vPorts and vSwitches. MTU inconsistency will cause packet loss and VM traffic forwarding performance degradation.

• Distributed FW Data Sync Detection: Checks whether the VM groups, IP object groups, and service object groups associated with the distributed firewall on a host have been updated. If the information is not updated in time, the distributed firewall rules on the host will fail to be synchronized. This can result in unsynchronized firewall policies, leaving the VM network exposed to potential security risks due to ineffective protection. Hosts running a lower software version (for example, a version earlier than E0920) do not support the distributed firewall data sync check item. Therefore, the health check result for such hosts is empty.

HA Service State

• HA Configuration: Checks whether VMs that have the same name exist on the source host after fault migration in a cluster.

• HA Service State: