We are constantly getting an error "host connection failure" for different ESXi host across the cluster and it last only for milliseconds, and there no significant impact in the enviroment. ESXi host events I can see the events "Host is not responding" all the "VM's disconnected" and the next moment ( I can say less that 1 sec) all "Vm's connected" back.
Short description about the environment:
vCenter version 5.0 and all the ESXi host's are v 5.0. We have 4 clusters out which 3 enabled with HA and DRS,and in one cluster not enable HA or DRS ,And each cluster have about 10 to 15 ESXi host.
Note :This issue not happening only for particular host, every time it happens to different ESXi host across cluster.
Any suggestion /direction to troubleshoot/find what actually causing the issue.
Thanks...........
Hi Pradeeshv,
What servers are you running ESXi5 on? Are these HP Blade servers? ALso check the NICs that are on the affected servers. Use ~ # esxcfg-nics -l to find out whether you have Emulex Corporation NC551i Dual Port FlexFabric 10Gb Adapter?
Thanks,
AG
Do all of your hosts share a common physical switch for their management interfaces? If so, I'd take a look at the switch to make sure it's healthy. Since your problem seems to affect all of your hosts, look for what they have in common.
Next I'd look at your vCenter host and verify that it's not havint connectivity issues itself.
Can you paste the vpxd logs when this issues happen.
The VM's are windows 2003 and windows 2008 servers ,all are cisco and IBM blade servers.And the NIC are
Broadcom corporation broadcom netxtreme ii BCM5708 1000base-SX
Broadcom corporation broadcom netxtreme ii BCM5709 1000base-T
Cisco systems inc cisco vic ethernet nic
Thanks...
No, Its not connected to the same physical switch ,and did n't find any network connectivity issues with vCenter server so far.
I couldn't find the vpxd log file ,when the issue happen,I could only see the last day vpxd log file,looks all the old log files are over written.And the issue happened 2 days back.
oh ok can you check if there is vpxd*.gz files archive .....
run the command ethtool -i <nic#>
run this for all the nics and let me know wats the driver nic has on it.
Hello Pradeeshv,
Check if there is a firmware upgrade and driver upgrade available for your servers. That does fix a lot of issues related to this.
Thanks-AG