Hi,
currently we hava two ESX servers in a HA/DRS cluster with 6 network interfaces. One interface is the service console. If I'm right, HA detects network failure, if the service console can't ping the gateway, respectively if the port goes down/is disconneted.
So what is the best practice? Do you have the service console designed redundant for HA purposes?
Thanks for support.
Regards
Marco
yes, yes, and yes if you are implementing HA (I also added vmotion to the service console standby nic to save a nic. This worked great for us.)
Hello,
Please remember that you really want to only use standby mode for vMotion and the SC as the load balancing methods are not going to work. Load balancing works by outgoing MAC, port, or IP and with the SC and vMotion there is only one outgoing MAC, port, or IP. Hence load balancing is not all that useful.
Also, if you are going to run vMotion over the SC network, please note that the SC network should be 100% firewalled from the outside... The reason for this is simple, in a failure mode, your VM memory images are sent cleartext over the SC (Administrative) Network. That would allow a hacker to access credential information stored within the memory image. If you can it is actually safer to have 2 more pNICs for redundancy of these items. It depends entirely on your level of trust in your Administrators, and the Administrative (Service Console) network.
Best regards,
Edward
Message was edited by:
Texiwill
Thanks for your reply.
In that case, I have another question: At this time, we have 6 network ports allocated. If we'd like to redundant the service console, we have to buy another network card. Perhaps we can consolidate the service console and the VMotion port, so we can save the second network card. But we have no experience, if this could overload the port or if this can produce further problems...
What is best practice in this case?
Thanks and regards
Marco
