October 16th, 2009 by Tim Phillips

Today, in the 3d party of my “In a nutshell” story I`ll talk about VmWare High Availability: what it is, and how does it work.

HA cluster can consist maximum of 32 servers, while 5 of them will be primary nodes. Primary nodes are keeping information about current configuration, and synchronize it between each other. If cluster includes less than 5 nodes, all of them will be primary.

By default hearbeat are sent primary-primary and secondary-primary every second, but you can change this parameter by command das.failuredetectioninterval.

First 5 nodes in cluster are primary, while all the rest – secondary. Both active, and passive servers are chosen randomly. The only way to see which servers are primary is to type the following command into service console:

cat /var/log/vmware/aam/aam_config_util_listnodes.log

or

/opt/vmware/aam/bin/Cli (ftcli on earlier versions)
AAM> ln
.

If one of primary hosts goes down, re-selection (one of secondary nodes becomes primary) doesn`t happen. It happens only if primary node goes to maintenance mode, if it`s disconnected or deleted from cluster.

One of primary hosts is Active Primary – it regulates reboot of VMs. If it goes down, another server will be assigned with this role. If all 5 primary servers go down - there will be no reboot of VMs, HA requires at least one active server for its work.

That`s all for today! Thanks and bye!

This website uses IntenseDebate comments, but they are not currently loaded because either your browser doesn't support JavaScript, or they didn't load fast enough.

Leave a Reply