We improved the Cluster Health tab. The quickest way to see these improvements is to watch a short shockwave demo . (No, it doesn’t have sound).
There are two big improvements.
1) Now the cluster names are color-coded so you can see at a glance which clusters have issues, just as the disk health is color-coded.
2) The other big improvement is that in addition to just seeing the current values of the cluster, we made it super simple to see how the cluster replicator has been performing over the course of the day simply by clicking on the server name. You could graph this information before, using the graph tab, but this is better we think.
Special thanks to Andy Pedisich’s Lotusphere presentation for pointing out that the trend is as important as the current values. He presented a manual way to get the type of information that VitalSigns gets automatically by using a modified statrep.nsf and Excel, which was pretty slick.
One interesting tidbit is that he mentioned was that in the course of his consulting he regularly encounters cluster servers that behind, and one was sometimes up to 90 minutes behind in the middle of the day, and without looking at these key metrics (which VitalSigns makes easy) most administrators don’t even know they have a problem until it is too late! One server crashes, the users fail over, and they ask “where did my last hour of work go?”
Thanks for a great presentation, Andy!
No responses yet