We just found out that when the SC cable on any ESX 3.0.1-server is disconnected from the switch and immediately is connected again, the SC does not respond to a ping for almost 40 seconds.
This means that HA kicks in when the network goes down for even a fraction of a second.
(To rule out the switch we connected a standard XP-laptop to the switch, pulled the cable and put the cable back immediately. Only 1 ping reply was lost.)
Question: is this normal behavior of the SC? Is there any thing we can configure to make the SC build up a network connection quicker and stop HA from starting?
Thanks in advance!
Regards,
Raymond
Try to hardcode the switch port and nic to 100mbit full/duplex and see if that helps.
Maybe the switch has spanning tree enabled? In this case a connect may take that long, unless you specify "spanning-tree mode portfast" (cisco).
We found the solution, the SC was connected to a trunk port with "spanning-tree portfast" enabled, but it should have been configured with the "spanning-tree portfast trunk" setting.
Thanks for your replies.
Regards,
Raymond