shankarvetrivel
asked on
Redundancy Managment network
Hello All,
I have a question about 'Redundancy Management network'
I am designing the network for new virtual infrastructure dedicated for a special project.
We are planning for 2 management networks for HA heart beats.
Vswitch 0
one VMkernel port group (used for management network)
It has 2 pnics drops to different physical switch.
IP address for Management network is
IP 168.127.133.10
Mask 255.255.255.0
GW-168.127.133.254
Vswitch1
one VMkernel port group (used for management network)
one VMkernel port group (used for VMotion)
It has 2 pnics drops to different physical switch.
IP address for second Management network is
168.127.134.10 Non routable VLAN
255.255.255.0
168.127.133.254
IP address for VMotion network is
168.127.134.11
255.255.255.0
Here is my question
Should I create DNS host entry for second management IP?
If not how HA will use secondary management network for Heart beat.
Thanks..
I have a question about 'Redundancy Management network'
I am designing the network for new virtual infrastructure dedicated for a special project.
We are planning for 2 management networks for HA heart beats.
Vswitch 0
one VMkernel port group (used for management network)
It has 2 pnics drops to different physical switch.
IP address for Management network is
IP 168.127.133.10
Mask 255.255.255.0
GW-168.127.133.254
Vswitch1
one VMkernel port group (used for management network)
one VMkernel port group (used for VMotion)
It has 2 pnics drops to different physical switch.
IP address for second Management network is
168.127.134.10 Non routable VLAN
255.255.255.0
168.127.133.254
IP address for VMotion network is
168.127.134.11
255.255.255.0
Here is my question
Should I create DNS host entry for second management IP?
If not how HA will use secondary management network for Heart beat.
Thanks..
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
when network isolation occurs which policy would be the best.
Shutdown Vm
Leave VM power on
Shutdown Vm
Leave VM power on
that depends upon your needs and situation.
With "leave powered on" you are saying...if the host loses connectivity from the service console to the isolation address (usually the gateway), I'm betting that the host has not lost the network connectivity for the VM's because I have them on another network and/or different network hardware. Thus, when the other hosts see the problem host go offline, they will try to power on the VM's that were running on that host and those power-on operations will fail and time-out because the files are locked by the isolated host.
The danger of this option is that you might leave the VM's powered on but they are not reachable from the network because the network failure was more far-reaching than you had guessed.
If you choose "power off" the danger is that all VM's will be powered off if you lose networking to all hosts and all of them think they are isolated. Should that happen, all VM's will remain powered off until you initiate power-on operations for all of them.
With "leave powered on" you are saying...if the host loses connectivity from the service console to the isolation address (usually the gateway), I'm betting that the host has not lost the network connectivity for the VM's because I have them on another network and/or different network hardware. Thus, when the other hosts see the problem host go offline, they will try to power on the VM's that were running on that host and those power-on operations will fail and time-out because the files are locked by the isolated host.
The danger of this option is that you might leave the VM's powered on but they are not reachable from the network because the network failure was more far-reaching than you had guessed.
If you choose "power off" the danger is that all VM's will be powered off if you lose networking to all hosts and all of them think they are isolated. Should that happen, all VM's will remain powered off until you initiate power-on operations for all of them.
ASKER
'Should that happen, all VM's will remain powered off until you initiate power-on operations for all of them'
Why we should initiate manual power on .I think HA primary agent will restart the VMs...
Please correct me...
Why we should initiate manual power on .I think HA primary agent will restart the VMs...
Please correct me...
No, if ALL of your ESX hosts have a network isolation event (i.e. they are all connected to only one physical switch and that switch crashes) then all of the hosts will think they are isolated and will power down their VM's in the belief that there are other hosts that aren't affected by the network failure. but since they all think they're isolated and there's no one host that thinks it is in a good position and should try to power on the VM's on the other hosts.
To protect yourself from such an event; you use multiple uplinks across multiple switches, but then if you lose power to all of your switch gear and your hosts are still up because they are on different power/UPS...the same thing will happen. So, then you gotta plan your UPS accordingly. and on it goes... :)
To protect yourself from such an event; you use multiple uplinks across multiple switches, but then if you lose power to all of your switch gear and your hosts are still up because they are on different power/UPS...the same thing will happen. So, then you gotta plan your UPS accordingly. and on it goes... :)
ASKER
Good
das.allownetwork will be the option you will want to look at.
http://www.yellow-bricks.com/2008/08/08/dasallownetwork-where-and-when/