VMware Cloud Community
zenomorph
Contributor
Contributor

Service Console/HA and vMotion redundency

Hi there....

I've read the numerous posts here on configuring redundency on the Service Console for HA heartbeat and also redundency for the vMotion network. However I'm still a bit confused after reading all the suggestions.

To put it straight we'll be using HA, DRS and vMotion our servers will have about 8 NICs some, I've estimated 4 for VM network and the remaining for other vSphere functions.

For redundency our network has a pair of Core switch for redundency and we have 3 vSphere Ent Plus. What I was thinking was the following for the each vSphere:

vSwitch0 - Service Console

NIC01- Active

NIC02 - Standby

(NIC01 and NIC02 will be connected to different Core switch for redundency should 1 core switch fail.)

vSwitch1 - vMotion

NIC3 - vMotion

I was planning to connect this to a dedicated Cisco 3650 switch justed to be used for the vMotion network.

Whatever NICs remaining will most likely be used for VM network.

My questions is;

Is this adequate redundency for the Service Console and HA use with 2NICs (active and standby no teaming)?

Do I need/recommend redundency on the vMotion network?

Cheers many thanks

All the NICs are gigabit

0 Kudos
3 Replies
AndreTheGiant
Immortal
Immortal

A good redoundancy solution is have at least 2 NICs on each vSwitch.

A good security and performance solution is to not mix different type of traffic on same vSwitch (or do it with VLAN functionality).

So you can consider to add another NIC to VMotion vSwitch, or merge VMotion and COS vSwitch and simple put different active/stanby rules for each portgroup.

See also: http://communities.vmware.com/docs/DOC-10985

Andre

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
zenomorph
Contributor
Contributor

Thanks for the tips.

That actually bring me to another two points I need some assistance to clarify:

1. most likely we'll team 4NICs to agrregate the throughput to 4GB with etherchannel and also VLAN tagging so that we can maintain the IP address of our servers on different VLANS/subnets without the need to change IP address. Are there any issues with VLAN tagging that may occur.

2. as we are using vSphere Enterprise it doesn't have the Network vMotion feature (port state follows VM) what are the issues of not having this on vSphere Enterprise i can't seem to find much information on it.

Cheers............

0 Kudos
AndreTheGiant
Immortal
Immortal

1. most likely we'll team 4NICs to agrregate the throughput to 4GB with etherchannel and also VLAN tagging so that we can maintain the IP address of our servers on different VLANS/subnets without the need to change IP address. Are there any issues with VLAN tagging that may occur.

VLAN is a good solution to isolate traffic.

Remember that Etherchannel require also to use the team policy: "IP hash"

2. as we are using vSphere Enterprise it doesn't have the Network vMotion feature (port state follows VM) what are the issues of not having this on vSphere Enterprise i can't seem to find much information on it.

Maybe you mean Essential?

If you do not have VMotion license you cannot have VMotion interfaces Smiley Wink

If you have VMotion license you do not use VMotion interfaces, then it's a pity

Andre

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro