Project HA

[Ivanildo_GalvÃ_]

Friends,

I am developing my first project with VMware HA, the client has 4 servers that will be done P2V them, the two hosts will provide Dell customers with two processors  and 32GB of RAM each.

The project is less expensive, as are just four hosts and VMS, will adopt the vSphere Essentials Plus Kit, perfectly fits the scenario, I will scale the part of the storage, I'm thinking to suggest this solution: http://bit.ly/m3kq13

Cost and benefit worth, I believe, goes with 8TB of Storage.

As a network, the client has two Gigabit Switches Dell, I will use them in the project, the network must be configured like this: One network to the console vCenter, One vMotion network and One LAN network of VMs ?

I can also deploy the solution vswitch distributed to ensure high network availability?

Thank you, wait!

[ranjitcool · Accepted solution]

Hello,

How well to design also depends on how large your design will scale, as in considering future growth as well.

For this project, I would stick with vswitches. Having distributed vswitches can be done after as well and to keep things simple, vswitches are great.

Best networking practises say have redudant management networks on different networks connected to separate swtiches.

Have separate vmotion network

Have separete storage network

Have separate vm network for vms to talk to.

Now you can do all the above with vlans easily instead of having a nic for each, this again depends on how much data you will be pushing or pulling through so run some stats on that physical box.

Rest I think you are fine. The iomega is good stuff.

R

Do mark correct answer if you liked it.

[Ivanildo_GalvÃ_]

Understand the customer's network is small, it will have 4 virtual machines at the beginning and is expected to increase to more than one or at most two virtual machines in the future, totaling six.

Therefore, the project will do so.

A - VMware vSphere Essentials Plus Kit because of the HA and VMotion.

B - 8TB of Storage Iomega with iSCSI LUN

C - 6 Veeam Backup & Replication for backup of virtual machines and can be used in future for replication of virtual machines to other VMware hosts.

D - E available on each server: A NIC for Storage, one for VMS and one forVMotion.

E - All NICs connected in gigabit switch that the client already has.

I will not design redundancy on each network link, as this generates more costs by buying more network card.

What do you think?

[ranjitcool]

Hello,

Before I answer your questions.

You have 4 vms

And how many physical hosts?

R

[Ivanildo_GalvÃ_]

02 Hosts and 6 VMS.

Thanks !

[MauroBonder]

Neste cenario fica dificil voce ter boa pratica junto a alta disponibilidade, pois se for pensar em alta disponibilidade seriam 2 placas de rede para service console, mais duas para vmotion, mais duas para Storage iscsi, mais 2 para as maquinas virtuais.

Complicado hein...rsrs

Agora o que voce pode fazer seria utilizar essas duas placas em teaming, que teras a redundancia e segmentar todas suas redes por VLAN para nao causar impacto uma rede na outra.

Seria uma vlan para Service console, outra para vmotion, uma para rede do Storage iSCSI e mais uma para rede servidores.

A outra opcao que nao teria a redundancia porem iria garantir a boa pratica pelo menos para rede iSCSI seria uma placa apenas para acesso do storage dedicado, e a outra placa para o restante, porem sempre com segmentacao por vlan.

Abcs!

Discussion moved from VMware ESXi 5 to Availability: HA & FT

[Ivanildo_GalvÃ_]

Olá Mauro,

Vamos deixar o termo Alta Disponibilidade de lado, o cliente quer apenas que as máquinas sejam reiniciadas automaticamente em outro host caso o host orginal das VMS venha a ter problemas.

No projeto que estou montando aqui, estou contemplando da seguinte forma.

São dois Hosts, Dell R410.

Cada host com 4 Nics Gigabit

Tava dividindo assim:

     01 Nic para o console

     01 Nic para as VMS

     01 Nic para o Storage Iomega

     01 Nic para o Vmotion

Além disso, o cliente ainda terá em outro Storage o backup de todas as VMS, para garantir aí um plano B, caso por exemplo o Storage das VMS apresente um sinistro e detone as VMS.

Pergunto.

Não poderia usar a mesma Nic do Storage para o Vmotion ?

Obrigado !

[ranjitcool]

Okay seriously, would be great if english was a medium here. Remember its an international language. :smileygrin:

[ranjitcool]

Hello,

I think you will be having a robust solution with what you have.

Like i said earlier, have a nic assigned for vmotion, management and lan traffic. In esxi5, you will have a to bind a nic to iscsi so be mind ful of that.

Veeam is a good product for backup and replication.

Update this topic if you have any other specific questions.

I think having both hosts in a cluster with the iomega storage is good.

[MauroBonder]

Sim pode! Mas crie um portgroup para vmotion e uma pro storage e de preferencia segmentadas por vlan.

Pro HA funcionar vc deve ter tudo configurado correramente nos dois hosts inclusive a parte de nomwnclatura de prtgroup e utilizar o storge shared pros dois hosts.

E a virada depende da prioridade q vc configurar por vm q por default eh medium pra todas e tenha certeza q um host pode segurar a carga de tudo q esta rodando atraves do admission control q vc pode ver no edit settiga

Do cluster

[Ivanildo_GalvÃ_]

É, pelo jeito tem que dar uma simulada boa e ler bastante os artigos pela Web pra saber botar esse negócio pra rodar.

Valeu pelo toque !

[psplnalin123]

First of all with vmware essential license distributed switch cannot be configure because it is not licensed with essential secondly right now you are having 2 NICs so, 1 nic you could dedicate to the managment trafic and vmotion and second nic you dedicate to virtual machines.

For more details you could refere and ask customer for Network redundancy and please refere below link for best practice of High availability

http://www.vmware.com/pdf/vsphere4/r41/vsp_41_availability.pdf refer page nbo . 28 & 29