20110922-vSphere5-ICM-Day4
VMware
Management PowerCLI for vSphere 5
VMware
vSphere Documentation (有行動裝置格式)
Resource
Pools
Expandable
Reservation
- 如果Reservation資源不夠可以跟上一層Pool 要資源
– class
break –
Monitoring
Resource Usage
Alam
Settings
- what to monitor
- Monitor for conditions or state
- Monitor for specific events occurring on this object
Notes:
- VMware 有提供 SNMP MIB for notification trap
Running
Nested VMs
monitor_control.restrict_backdoor = TRUE
VMware
ESXi 5.0 可以在 Nested
VM 上面run
64 bits OS
Notes:
- VM 如果要換Storage,
在VM
上面按滑鼠右鍵,
Migrate
- 進行 Storage Motion 點選自己的ESXi Host
- 再執行目的地 storage 的選擇
** Module 11 High
Availability and Fault Tolerance **
vSphere HA
Failure Scenarios
- ESXi host failure
- When a host fails,
vSphere HA restart the affected virtual machines on other
hosts.
- 重新啟動原來 fail Host 上的 VM(而非vMotion) 在其他的Host
- When a host fails,
vSphere HA restart the affected virtual machines on other
hosts.
- Guest OS failure
- vSphere HA resets the
virtual machine
- vSphere HA resets the
virtual machine
- Application failure
- When an application
fails, vSphere HA restart the affected virtual machines on the
same hosts.
- Require VMware Tools to
be installed
- When an application
fails, vSphere HA restart the affected virtual machines on the
same hosts.
Notes:
- Distributed Resource
Scheduler ( DRS) 自動會做 vMotion
Configuring vSphere HA Settings
- Host Monitoring Status
- Admission Control
- 是否要預留資源給指定的Host
- Admission Control Policy
- 留多少資源給指定的Host
- Host failures the cluster tolerates
- 使用 Slot size ( 所有 Power VM 的CPU / MEM 資源的最大值) 觀念來保留資源
- 例如總共資源為 CPU = 8G / MEM = 16G, 一個VM 要使用到 CPU 2G, MEM 8G 那一個Slot 就是只能放兩個VM
- 可以在 Cluster 的 summary Tab, Adanced Runtime Info 內觀察 Slot size
- Percentage of cluster resources as failover spare capacity
- Specify failover hosts
- VMware 建議使用 Percentage 的方式
– class break –
vSphere High
Availablity Architecture
vSphere HA
Architecture
- vSphere 4.x 使用
AAM
- vSphere 5.0 使用
Fault Domain Manage
- 連接 Datastore 最多的 host 當master
- vCenter Server 負責設定, ESXi Master 負責監控
Notes:
- 如果 HA Master Fail, 會從 Slave 挑出連接最多 datastore 來繼任Master, 如果 Slave 的datastore數量都一樣, 那就取 MOID (Managed object ID) 數字第一位數最高來當 Master
- vSphere HA provide a base level of protection. FT ( Fault-tolerant) provide a higher level of availability. ( FT 把VM clone 到另外一台 ESXi Host 並保持同步)
- FT 必須基於
vSphere HA 架構下,
只能做到 1
個VCPU,
磁碟要使用 Thick
Provision Eager Zeroed
- VM 如果設定 FT, CPU 及記憶體會調到最高, 會影響到 Slot 的數量
elearningModule
on Fault Tolerance Demo
本尊在這邊
網路上找到的相關資源
Greetings, As you
may be aware, the following eLearning Demos are available for
situation where you cannot perform the lab exercises, such as due to
lab environment limitations, including lack of server VT support etc:
eLearning Module
on ESXi Installation and Configuration
http://mylearn.vmware.com/register.cfm?course=38258
eLearning Module
on Performing a Cold Clone
http://mylearn.vmware.com/register.cfm?course=38260
eLearning Module
on Performing a Guided Consolidation
http://mylearn.vmware.com/register.cfm?course=38265
eLearning Module
on Enhanced VMotion Compatibility
http://mylearn.vmware.com/register.cfm?course=73277
elearningModule
on Fault Tolerance
http://mylearn.vmware.com/register.cfm?course=73276
elearningModule
on VMware Data Recovery
http://mylearn.vmware.com/register.cfm?course=80063
Though these
links are already posted in the Trainer's Community, to make it
easier for you to provide these links to your students, the links are
also now posted on the vdcstudents.com web site at:
Username: student Password: script39
Lab Notes:
- 如果把 Cluster 移除掉,
要重新把所有原本在 datacenter
內的主機再次加回來,
但是不影響運作
沒有留言:
張貼留言