It has been a while since I made the time to write but I found this update on vCSHB very interesting and worth sharing. I have been testing vCSHB 6.3 in the lab with a stretchd datacenter and ESXi cluster design for about a year now and found it almost impossible to manage vCSHB given it’s reliance on a singe host IP/Name. Mike Laverick had a great post on the product back in August of 2010. Since that time VMware has been working on improving this fundamental issue and managed to slip it into a minor update. I suspect most people missed this “minor detail” in the 6.3 U1 release notes. It reads:
“Enhanced passive server management capabilities — A new deployment option allows the passive server to be managed and monitored remotely, this includes receiving file level antivirus updates…”
THIS IS A BIG DEAL! This allows us to finally use a third IP to manage the vCenter instance while providing both HA nodes their own (unique) primary host IP. What a relief. I am very happy to report that this has been the best release of vCSHB I have seen yet and with this feature, it is finally something I would consider putting into production when the time is right.
I included a high level overview below that may help explain:
What may not be entirely obvious here is what to do with the Neverfail IP filter driver. Why do we still need a filter driver controlling ingress and egress traffic on primary interfaces for the vCenter hosts if we now have the VIP? VMware ensures me it still plays a part but I can tell you we have it turned off and haven’t noticed a problem yet! :-)
Enjoy!
