Module 9: Deploying and Configuring vSphere Clusters PDF

Module 9 Deploying and Configuring vSphere Clusters 9-2 Importance Most organizations rely on computer-based services such as email, databases, and web-based applications. The failure of these services can mean lost productivity and revenue. By understanding and using vSphere HA, you can...

Module 9 Deploying and Configuring vSphere Clusters 9-2 Importance Most organizations rely on computer-based services such as email, databases, and web-based applications. The failure of these services can mean lost productivity and revenue. By understanding and using vSphere HA, you can configure highly available, computer-based services, which are important for an organization to remain competitive in contemporary business environments. And by developing skills in using vSphere DRS, you can improve service levels by guaranteeing appropriate resources to virtual machines. 9-3 Module Lessons 1. vSphere Clusters Overview 2. vSphere Distributed Resource Scheduler 3. Introduction to vSphere High Availability 4. vSphere High Availability Architecture 5. Configuring vSphere High Availability 439 9-4 Lesson 1: vSphere Clusters Overview 9-5 Learner Objectives Create a vSphere cluster Recognize cluster options that you can configure with Cluster Ouickstart View information about a vSphere cluster 440 9-6 About vSphere Clusters A cluster is used in vSphere to share physical resources between a group of ESXi hosts. vCenter manages cluster resources as a single pool of resources. You can create one or more clusters based on the purpose each cluster must fulfill, for example: Management Production Compute A cluster can contain up to 96 ESXi hosts. vSphere Client 0, [hJJ ~ sa-vcsa-01 vclass local v Ell SA-Datacenter > [{JJ SA-Compute-01 v [[;]] SA-compute-02 ld sa-esxi-04.vclass.local ld sa-esxi-OS.vclass.local !:] sa-esxi-06.vclass.local ~ Linux01 Gfi linux02 Gfi Photon-01 fti Photon-02 E~ Photon-03 > [(Jj SA-Management In this screenshot, five clusters are shown: SA-Compute-01, SA-Compute-02, SA-Management, SB-Development, and SB-Management. To support 96 hosts per cluster, vSphere 7 Update 1 or later is required. 441 9-7 Creating a vSphere Cluster You can create a cluster by giving it a name and selecting the relevant cluster services. New Cluster I·· I t· Ld!l S!:tYe-1'5.:i1 SA·CCr;"'pU!e·01 ~-V(SD ·OI ;(.t.llstb:al You can activate the following services in a vSphere cluster: vSphere HA, for high availability - vSphere DRS, for VM placement and load balancing vSAN, for software-defined storage You can also manage host updates using images. With vSphere Lifecycle Manager, you can update all hosts in the cluster collectively, using a specified ES Xi image. 442 9-8 About Cluster Ouickstart After you create a cluster, you can use the Cluster Ouickstart workflow to configure the cluster. = vSoh€1€ C-. 1 t, - rf.H !CM-Compute-02 : J.-H'""~ IJ' IC.M·OJt·Oi !.Cluster bCI· 01\ WS SETT INGS I pr101ity rn.::m ty 3 2 ;:;nd ti:(CJr"l"T;7'1cf,;t:O'"if. To view the vSphere DRS pane, go to the cluster's Summary tab. 464 9-30 vSphere DRS Settings: VM-Level Automation You can customize the automation level for individual VMs in a cluster to override the automation level set on the entire cluster. ltD ICM·Compute·02 ; ''"o"' s.rv~c:u.., VM Overrides Add VM Override Add VM Override VM.'H MI Ru!~~ IEM vSphNt'! HA VM R:~Slilll f>IICflty ,::.dOi\,oM1 c:.;,iay: CJ o~·Umg!> f,l'lc~ Re~P'>'lSe t!) l] ~ ll.:fe uo.. e.ou!'f~· !]CWt'trlU'!': vSphlo'.re HA- VM ~1oni\oang By setting the automation level for individual VMs, you can fine-tune automation to suit your needs. For example, you might have a VM that is especially critical to your business. You want more control over its placement, so you set its automation level to Manual. If a VM's automation level is set to Disabled, vSphere DRS does not migrate the VM, nor does it provide initial placement or migration recommendations. Select the automation level based on your environment and level of comfort. For example, if you are new to vSphere DRS clusters. you might select Partially Automated because you want control over the movement of VMs. When you are comfortable with what vSphere DRS does and how it works, you might set the automation level to Fully Automated. You can set the automation level to Manual on VMs over which you want more control, such as your business-critical VMs. 465  9-31 vSphere DRS Settings: VM Swap File Location ESXi hosts can be configured to place VM swap files o n a local datastore. If vSphere DRS is configured , you should place the VM swap files on a shared datastore. I l !Cll ICM-Compute-02 : ~r1·0t; ~ l :i '..lrr' lllill)o M;:oftr !Ct C. 1 f ~·.Ire f:O:IITl!~~IC!'l~ H,;l~l!. V Mr (!o, :;,;;tO!'"o:,; l General I ~. ::·::.~:~::."~*fi SW~P ~IOC~liOn Default VM Compalihilil )' V l~lUd[ mi.Cnll'l@ d l!~C\O t y Thrs is !he def"l !~ tompitt lb~i! y lor...-irtu:~l ;na~hrn~ oea Canflgllrauon Ed it Clu ster Settings ICM-Compute-02 X Camphty Contlguro)rion Crea te VM/Host Rule ICM-Compute-02 X KO!'EVC l BalanceWorio:loads ' '"1'"111 (1.. ' " 'H l"" ' 1/MCtvo!Hidl:'c$.J Enable rule. 1/0f'ii!N!'i ····--·1 Type I Jscparare VIrtual M£~chi~t:s Keep Virtuat Machines Together Tile listed \hrtual v.ach!nO:::!> must Separate Virtual Machines ADD. REMOVE Virtual Machines to Hosts Virtual Machines to Virtual Machines I 0 [) ).t'!'tll~ r t. (~ Wtn1C"1·04 0 (~ Wml0.. 06 !II After a vSphere DRS cluster is created , you can edit its properties to create rules that specify affinity. The following types of rules can be created: Affinity rules: vSphere DRS keeps certain VMs together on the same host (for example, for performance reasons). Anti-affinity rules: vSphere DRS ensures that certain VMs are placed on different hosts (for example, for availability reasons). If two rules conflict, you are prevented from activating both. When you add or edit a rule , and the cluster is immediately in violation of that rule, the cluster continues to operate and tries to correct the violation. For vSphere DRS clusters that hav e an automation level of manual or partially automated, migration recommendations are based on both rule fulfillment and load balancing _ 467 9-33 vSphere DRS Settings: DRS Groups VM groups and host groups are used in defining VM- Host affinity rules. The VM- Host affinity rul e specifies w hether a VM can or cannot be run on a host. Types of groups: VM group: One or more VMs Host group: One or more ESXi hosts A VM can belong to multiple VM groups. A host can belong to multiple host groups. rt'JJ ICM-Compute-02 ; ' "'o" "'·''' !!..- 1·"1! ~· ·.'~-~~.., ~~"'· r;!r,ur.l :.:.~~-:,·e::: I Create VM/Host Group ICM.C omp>.:tB-02 · v·-v -xllll ,""".'" -~ For ease of administration, virtual mac hines can be placed in VM or host groups. You can create one or more VM groups in a vSphere DR S cluster, each consisting of one or more VMs. A host group consists of one or more ES Xi hosts. The main use of VM groups and host groups is to help in defining the VM-Host affinity rules. 468 9-34 vSphere DRS Settings: VM-Host Affinity Rules A VM-Host affinity rule: Defines an affinity (or anti-affinity) relationship bet ween a VM group and a host group Is either a required rule or a preferential rule Rule options: Must run on hosts in group Should run on hosts in group Must not run on hosts in group Should not run on hosts in group 1h11 ICM-Compute -02 : At liONS SNvlces v VM/Host Rules Conflgur:Jtlon Create VM/Host Rule ICM -Compute-02 X Vt.1W Oa lastor~ fa r H D.l t.U!Or~ wtfl'l A PD ·t,: ' '( I ;.,,,.. ,.,., · ;,!' V M M oolt o rl ng In the vSphere Client. you can configure the following vSphere HA settings: Availability failure conditions and responses: Provide settings for host failure responses, host isolation, VM monitoring, and VMCP. Admission control: Activate or deactivate admission control for the vSphere HA cluster and select a policy for how it is enforced. Heartbeat datastores: Specify preferences for the datastores that vSphere HA uses for datastore heartbeating. Advanced options: Customize vSphere HA behavior by setting advanced options. 508 9-73 vSphere HA Settings: Failures and Responses You use the Failures and responses pane to configure a cluster's response if a failure occurs. Edit Cluster Settings ICM-CompUt€-02 X vSph€-t~ H;. t[) I F ·;irurc~ :·md re!.ponse~, I Adrmssron Conuol Heartbeat Datastorc.s Advdnced Options You can configute how vSphere HA responds to the falfUJe conditron.s on this cluster. The following failure conditions are supported: host. host isolatron. VM componen t protecllon (datastore wrth POL and APD}. VM and application. > Host Failure Response > Response tor Host Isolation > Datastore wllf'l POL > Datastore with APD Diw bled > VM Monitoring Power off and resto1rt VM5 · Conservatio,e restart policy Power off and fl?:Start VMs Agoressive rest.art poticy Using the Failures and Responses pane, you can configure how your cluster must function when problems occur. You can specify the vSphere HA cluster's response for host failures and isolation. You can configure the VMCP actions when permanent device loss and all paths down situations occur and activate the VM monitoring. The following host failure responses are available: Disabled: Host monitoring is turned off and VMs are not restarted. Restart VMs: VMs are failed over based on their restart priority. The following responses to host isolation are available: Disabled Power off and restart VMs Shut down and restart VMs 509 The following responses to a datastore PDL condition are available: Disabled Issue events: No action is taken against the affected VMs. The administrator is notified when a PDL event occurs. Power off and restart VMs The following responses to a datastore APD condition are available: Disabled Issue events: No action is taken against the affected VMs. The administrator is notified when an APD event occurs. Power off and restart VMs - Conservative restart policy: vSphere HA does not attempt to restart the affected VMs unless vSphere HA determines that another host can restart the VMs. The host experiencing the APD communicates with the primary host to determine whether sufficient capacity exists in the cluster to power on the affected VMs. If the primary host determines that sufficient capacity exists, the host experiencing the APD stops the VMs so that the VMs can be restarted on a healthy host. If the host experiencing the APD cannot communicate with the primary host, the VM is not stopped. Power off and restart VMs - Aggressive restart policy: vSphere HA stops the affected VMs even if it cannot determine that another host can restart the VMs. The host experiencing the APD attempts to communicate with the primary host to determine if sufficient capacity exists in the cluster to power on the affected VMs. If the primary host is not reachable, sufficient capacity for restarting the VMs is unknown. In this scenario, the host takes the risk and stops the VMs so that they can be restarted on the remaining healthy hosts. However, if sufficient capacity is not available, vSphere HA might not be able to recover all the affected VMs. This result is common in a network partition scenario in which a host cannot communicate with the primary host. The following VM monitoring options are available: VM monitoring only VM and application monitoring 510 9-74 vSphere HA Setting: Default VM Restart Priority The VM restart priority determines the order in which vSphere HA restarts VMs on a running host. VMs are assigned the Medium restart priority by default, unless the restart priority is explicitly set using VM overrides. Edit Cluster Settings ICM-Compute-02 X vS;:her~ HAC) F.:t1lu:es and re~ponses AdmJSS!on Conuol Htartbeiit Oatastor;.:-,nnortrq G) C) "' Host ~.allure Qesponse Failure Pesponse Allows you to conr,gure host mcmton11g and fa1tover o... thi~ S ~-oliT:.." !~. :£1 J~ Heartb~::.ats dete,red.....:.:.:.::..c..:.::..::....:.cc:.:::.::...__ _ \ Gues.t Hea rtbeat'j d~h!tt~d Re50urc~s attocated Powered On After a host failure, VMs are assigned to other hosts with unreserved capacity, w ith the highest priority VMs placed first. The process continues to those VMs with lower priority until all have been placed , or no more cluster capacity is available to meet the reservations or memory overhead of the VMs. A host then restarts the VMs assigned to it in priority order. If insufficient resources exist , vSphere HA waits for more unreserved capacity to become available, for example, because of a host coming back online. and then retries the placement of these VMs. To reduce the chance of this situation occurring, configure vSphere HA admission control to reserve more resources for failures. With admission control, you can control the amount of cluster capacity that is reserved by VMs, which is unavailable to meet the reservations and memory overhead of other VMs if a fai lure occurs. Optionally, you can configure a delay when a certain restart condition is met. 511 The following cond itions must be met before a VM is considered ready: VM has resources allocated VM is powered on VMware Tools heartbeat is detected VMware Tools application heartbeat is detected 512 9-75 vSphere HA Settings: VM-Level Restart Priority You can customize the restart priority for individual VMs in a cluster to override the default level set for the entire cluster. ;o; ICM-Compute-02 ; ,...,,""' Se!vlc~,..... V M Oven ides Cooflgu d Addil!o r.al d e l"i'" Q C¥1'U;,J~ lOIIh'~t \IM rt:~U-H pr io;ity con..1rhon 0 O ~mte HA.. POL Protection Sctt:ng5..____ vSpllertc>HA - APD P1 ot~chon Sel!ir1gs Fo~:lure Rcsponre (i} [J o-.·t'fr;