I have a two host ESXi 5 cluster and for some reason my virtual machines lose their connection to the internet every night. LAN connection works just fine during this time. As far as I know this only applies to our VMs.
Any one have any ideas on this?
thanks for the update. dont forget to make answer as useful or correct
Additionally. The connection problem only lasts for a few minutes and then things return to normal. Normally this would't be a real problem, but one of the servers needs connectivity for validation purposes.
check in alrams section for the same. Also cechk for event logs within vm to see if application works or hangs during tht time. this is to narrow down whether issue is with network or underlying storage on which vms are hosted
Thank you for your response. The VMs event logs don't register anything of interest...only a few applications register connectivity loss. VMs on either host have the same problem and the hosts don't show any alarms.
can u also check logs on the storage?
IBM Storage event logs don't show any recurring events that correspond to the connectivity problem. Storage errors seem an unlikely cause for this problem since the VMs fuction normally in other respects.
one last question, is it happening on all vm's on all esxi servers ?
also all esx connected to same physical switch
I will verify this on a few more servers tonight. But as far as I know this happens on all VMs on both hosts. The VM networks associated with the VMs have two physical network adapters and are connected to 2 different switches. The same switches have other servers(hardware) connected to them and they don't lose connectivity (will verify this again tonight).
kul..simple reason to ask was to narrow down the. issue to virtual or physical layer .. if it is happening on both servers, it is less likely to be config issue. could be a network dip on physical switches
Hi. Does it happes always at the same time? R u running something at that time? Backups or other things that might cause this?
Thank you for your answers. After doing some more tests it seems the problem isn't related to the server infrastructure. Sorry for the hasty post.
thanks for the update. dont forget to make answer as useful or correct