Some 2 Second Down Time is faced in Failover in VMWare

[SAKI201110141]

Hi everyone,

I have configured VMware for FT. I have created one VM (one CPU) with windows 2003 Server R2 with SP2. I am running IntelliMAX (HMI/SCADA application) on that VM. When I failover any ESX Server, then a distortion in Trend data is observed at that time of failover for around 2 seconds. Also I face 3 seconds dataloss and downtime on failback during VM synchronization. I have allocated 3 GB RAM to my VM. Also I am using 1GB NICs and GigaBit Switches. I am using StarWind as iSCSI SAN with redundant Storage links. I have used one FT link, one VM/Service Console link, and 2 Storage links on each ESX server.

For ESX Server, I am using Dell Optiplex R710 Servers with 8 GB RAM, PERC H700 RAID Integrated Controller Card, and E5620 Quad Core processor.

I also want to ask, whether we can use multiple cores for one VM in FT environment in VMWare. I am using ESX4.0 and vCenterServer 4.0. Also, as IntelliMAX is a web based application, and its client can be launched from any node on network. Is there any way in VMWare, that we can configure alarms to be displayed at the Application's node when any ESX server is down.

[ddecker]

Only 1 cpu per node in FT. I have only dabbled with but no it is very network intensive - are you using trunked lines shared with vmotion or other vmkernel processing.

As for when a node is donw you can setup alarms from any object in the hiarchy you want - the easiest is mail - if you do that remember you have to go to Administration and  Vcenter Server Settings and setup SMTP. I attached an example for you.

[SAKI201110141]

Hi ddecker,

Thanks for your reply. As I am a new user of VMWare, so I am sending you the snapshots of the configurations which I have performed on vCenter Server and my topology as well for reference. Please look into the topology and configurations, and tell me if there is a mistake in the configurations or network topology.

Regards.

SAKI