vsphere High Availability

The physical machine outage virtual machine switches to another physical machine to continue running. If the virtual machine freezes, then the virtual machine is reconfigured with high availability.

as we have learned before STP ( Spanning Tree Protocol ) also enables highly available clusters

aspects of High availability implementation: (mainly divided into 4 areas)

1, the application layer of high availability: the implementation of SQL Server,MySQL and other service programs cluster, to determine whether the application is running frequently.

2, operating system high availability: to the operating system cluster, operating system is the platform to run services, mainly to its service program to do high availability.

3.High availability of the virtualization layer:HA,FT minimizes or eliminates unplanned downtime by providing fast recovery and continuous availability of interrupts .

4, the physical layer of high availability: network card, storage redundancy

using vsphere , you can:

1, independent of hardware, operating system, and applications to provide higher availability. No special configuration is required.

2. Reduce planned downtime for common maintenance operations

3, in the event of failure to provide automatic recovery.

There are a number of instances like solving a single point of failure:

Ethernet Channel (switch)

HSRP (Hot Backup Routing protocol routers)

vsphere High Availability (Virtualization)

First, VSphere HA provides fast interrupt recovery

Cluster (a collection of multiple hosts, which requires at least two hosts)

The host cluster is implemented to provide fast recovery and cost-effective high availability for applications running in virtual machines.

HA achieves high availability in the following ways:

• prevent server failures-------physical hosts by restarting virtual machines on other hosts within the cluster

• Prevent application failures by continuously monitoring virtual machines ( sending detection messages via Vmtools) and reconfiguring them in the event of a failure. ------- virtual machines

features of Vsphereha:

1.No special software is installed in the application or virtual machine, and all workloads are automatically protected by HA .

2. HA and DRS are combined to prevent failures, and load balancing is achieved within the cluster.

DRS distributed resource Scheduling and distribution program

An HA cluster allows the ESXI host collection to work together as a group, which is higher than the level provided by a single host.

The hosts in the cluster are monitored and, if a failure occurs, the virtual machines on the failed host are restarted on the standby host.

There are two roles in a cluster:

Preferred Host Master

Slave host slave

convex ^-^ convex with DR and bdrin the OSPF we learned before 0> Similar, no preemptive right

Criteria for the campaign's preferred host:

• hosts in the cluster participate in the campaign (except in standby, maintenance mode, disconnected hosts), depending on the number of mounted data stores. ------'s a big possibility .

• Who joins the cluster first

there is only one preferred host in each cluster, and all other hosts are subordinate hosts. If master fails, shuts down, or is removed from the cluster, a new election is made.

However, the workload of the preferred host is also large:

2. Monitor the status of the subordinate host, and if it fails or is unreachable,master determines which virtual machine needs to be restarted.

4. Monitor all protected virtual machines (virtual machine status is on ) the power condition.

6. Manages the list of clustered hosts and protected virtual machines, maintaining inventory within the cluster.

8. Cache cluster(cluster) configuration,Master notification alerts the slave host.

10. The master host sends a heartbeat message to the slave host, and if the slave host does not receive a heartbeat message, re-election to the new preferred host.

12. Master reports status information to vcenter,and vcenter normally communicates only with master hosts. Virtual machine protection to ensure that the virtual machine is restarted when it fails.

instead, the subordinate hosts have only 3 tasks :

1. monitors the state of the virtual machine running locally and sends changes to the master host.

2. Monitor The health status of master host and if the host fails,slave participates in Master 's campaign

3. Slave uses HA access control function, does not need master 's coordination.

Access control:

there are sufficient resources within the cluster ( CPU , memory resources) for failover

And make sure that the current virtual machine resource is available.

Host failure type and detection

Master and the slave send heartbeat messages to each other.

in the HA cluster, three types of host failures are detected:

• Host stopped running (failed)

• host and network isolation ----- Network Isolation

• host loses network connection to preferred host ----- Network Partition

Detection path: Can be managed by the network, storage network

when Master cannot be contacted through the management network slave , it does not necessarily indicate slave Something's wrong, Master is checked through the storage network. slave is still alive.

vsphere High Availability