I have ESXi 6 set up on a small lab environment of three computers, including the visor. I am connecting directly to the host via IP (192.168.0.230) from a vSphere 6 desktop client. Note that I am not using vCenter Server in my environment, this is a direct connection to the ESXi host. This is a new ESXi 6 setup with no custom configuration hosting only a couple VMs, as vanilla and basic of a setup as you can get.
From the first time I tried connecting to the visor I had sporadic connection issues. When attempting to connect to the visor, 9/10 times it does not connect and gives me this error:
"vSphere Client could not connect to 192.168.0.230. An unknown connection error occurred (the request failed because the remote server took too long to respond). (The operation has timed out.)"
However occasionally it will log me in without a problem (1/10 tries or so), and I can do various configuration tasks for a while before the client finally crashes with a generic Windows error "Vpx client has stopped working".
I can ping the visor from either client without a problem, and I can access the visor's web page by typing its IP address into my web browser from either client as well. Network connectivity does not seem to be an issue, even among the VMs on the host.
When the VMs are running I can still access them without a problem even when I can't access the visor.
I have tried:
- restarting management network
- restarting all management agents from the DCUI
- installing vSphere client on another workstation
I have enclosed a snapshot of the DCUI log, it has one or two warnings I don't understand.
Any ideas?
How many NICs on the Management network?
Try paring that number down to just a single NIC if there are more than one NIC assigned. Being ambitious right out of the gate and assigning multiple adapters results in disconnects due to the default policy "route based on originating port ID)
How does the network test look in DCUI?
If you open the shell and logon as root, can you ping the gateway?
Only one NIC on the visor, which is a Dell Optiplex 980. I've used this exact hardware with ESXi 5.0 for a couple years and never had any problems. This is a simple lab environment, one visor connected to two Windows 7 x64 Pro workstations, each running vSphere 6 desktop client. They are connected to each other through a gigabit switch, no other external connections to the outside world. Because of this, the assigned gateway (192.168.0.254) is not a valid address, and neither are the two designated DNS servers, but ESXi forces you to input something. I am trying to connect directly by IP address, and about 1/10 tries it actually connects. I am able to do normal configuration tasks for 15-20 minutes, then the desktop client crashes in Windows with a generic error "Vpxclient stopped working". This only applies to the first client, the second computer with the vSphere 6 client has never been able to connect to the visor at all, I always get a timeout error. By all other accounts the visor is running fine, it responds to pings, I can browse to the visor's website in IE, and any running VMs continue to run normally and are accessible.
I'm having a similar problem. My esxi 6 is
[root@:~] uname -a
VMkernel esxi01 6.0.0 #1 SMP Release build-3073146 Sep 16 2015 04:08:47 x86_64 x86_64 x86_64 ESXi
My client is Windows 7, fully up to date on patches as of a few days ago.
I am able to log in to the server with ssh with no issues. The ESXi firewall is disabled. I am on the same subnet as the ESXi server.
Restarting the management agents from the console seems to have had no effect.
Hardware is Dell M610, which is on the HCL. Dell firmware is up to date as of 30 days ago. I used the ESXi ISO customized for Dell to do the install. Connecting with the client usually times out. Sometimes I get a certificate warning; when i accept the cert, i immediately get the timeout. Sometimes, it just times out.
FYI, I started having this problem with only one Windows 7 computer after I upgraded a customer ESXi server from 5.0 to 6.0. My vSphere clients I used every day on my laptop worked fine when on-site, but this one client is the one I use remotely to access the ESXi server on a regular basis and it stopped working, giving this error, after the upgrade (it had been working for about 4 years prior to the upgrade).
I tried restarting the vpxa and hostd services on ESXi, but that didn't work and obviously couldn't be the problem if I could connect from the same vSphere client version on my laptop. I tried uninstalling the 5.0, 5.1, and 6.0 clients from the problem computer, rebooting, then reinstalling the 6.0 client, but that didn't work.
The response to check if you can ping the router from ESXi led to the fix: I could ping the router from ESXi no problem, then verified I was using the same router on the problem computer. When looking at ipconfig, I noticed IPv6 was on, so decided to go turn it off since it causes random problems at other customers (AD and DNS interference).
After turning off IPv6 on the problem computer, the vSphere 6.0 client connected just fine.
This could be because I turn off IPv6 on ESXi after install to avoid any possible problems like I've seen at other customers, but the most important thing is that having IPv6 turned on on the problem computer was the cause of this error in my case.
