vMotion
- vMotion is a migration process
- vMotion allows you to migrate a running VM from one ESXi host to another.
- Can be used to load balance ESXi hosts in a cluster by moving VMs to a host with more available resources.
- Allows you to perform hardware adjustments by moving all VMs to another host
How does this work
- VMs are built on virtual hardware and are predominantly a large chunk of memory on an ESXi host
- Just need to copy that memory from one host to another
- Migration is very fault tolerant. If something goes wrong, the original host remains in control.
- Can operate without shared storage – can migrate even if the VMs do not use shared storage.
Distributed Resource Scheduler (DRS)
- Monitors memory and CPU loads on ESXi hosts.
- Upon VM Power up, DRS can recommend which Host to run the VM on, OR:
- Using vMotion, can automate migration of VMs to another host that isn’t being as utilized based on usage and configuration settings (rules).
- Using rules, some VMs can be bound together or forced apart (redundant clusters).
- Upon VM Powerup, DRS can recommend which Host to run the VM on.
Distributed Power Management (DPM)
- Leverages the automation of DRS to consolidate VMs on specific hosts, allowing low-use hosts to shut down and save power
- Power can be restored using Wake On Lan or IPMI
- On Power-up, DPM again leverages DRS to balance the ESXi hosts
Storage vMotion
- Allows migration of a VMs hard drive from storage device to another.
- Configures ESXi host to see both storage array, then I/Os are written to both source and mirrored disks.
- Storage devices do not need to be the same, as long as they are both supported.
- Can move VM from Shared device to Local Storage to perform maintenance on Shared while not disrupting the VM
Storage DRS
- Load balances Disk I/O between storage units
- Can be manually or automated
Storage I/O Control
- Some applications may consume all I/O available, at the expense of others (Database Servers, busyEmail Servers)
- Storage I/O Control makes sure some I/O is available for the lesser consuming applications to optimize performance.
vSphere Data Protection (VDP)
- Image Based Backup (aka Bare Metal Backup)
- Can be restored as a whole and powered on (like a snapshot) or
- Can be used to selectively restore specific files.
- Leverages data duplication to make backups efficient
- VADP provides open set of APIs to 3rd party backup providers so probably supported by your current backup software
High Availability (HA)
- Helps protect against hardware failures
- vCenter used to configure, but is not required to operate – each host participates in a “monitoring relationship”
- When a Host fails, the VMs will be migrated to the surviving Hosts and restarted. The VMs will crash, but are often restarted with a minute or so.
- Special software and drivers can also detect a VM failure. If the OS crashes or other instability is detected, HA can restart the VM automatically.
- By leveraging part of VMwares “Hyperic Product Suite”, HA can monitor specific services and applications even if the OS remains active.
Monitor ESXi Hosts
- Configured by vCenter, but do not rely on it to monitor hosts
- Each ESXi host participates in monitoring other hosts
- If a host crashes, the VMs are distributed between surviving hosts
- * The crashed VMs will need to restart, but are usually up within a minute or less.
Monitor VMs
- Can also be used to monitor VMs using a special set of software and drivers. These talk to the ESXi host
- Should this communication fail due to OS fault or other issue, HA can restart the VM automatically.
- Highly configurable, so different criteria can be applied.
Monitor Applications
- Leverages WMwares “Hyperic product suite”, can monitor applications and services.
Fault Tolerance (FT)
- Used for applications that MUST NEVER FAIL, or can cause data corruption if they do
- Perpetually mirrors specific VMs to another ESXi host
- Used “active-passive” clustering that is completely transparent to the application
- Once a failure occurs, the “passive” VM becomes the primary and a new secondary is created automatically.
- Easy to configure
- FA VMs are limited to 1 CPU in 5.x and up to 4 CPUs in 6.x depending on licensing.
vSphere Replication (VR)
- Used for Data Center Replication (between data centers, different sites, etc.)
- Can also be used to replicate data to a central location for data protection or recovery. (Remote locations – > Corporate office)
- Can be restored at a “point in time” in case recent changes are corrupt.
- Similar to R1Soft, you configure replication interval at remote site.
- Can monitor replication to ensure it is working.