Hi - can anyone help with this please
I've been using Vsphere 4 for a couple of years at my last company - no problems, bullet proof reliaility
Recently changed jobs and my current company is now starting to virtualise so I'm trying to convince them that VMware is the path to take. As a test to demo that there are no problems I have the free version of ESXi5 up and running - patched to build 768111. One guest VM running Windows Web Server 2008 R2.
All ran OK for about a week, then suddenly started getting
unable to connect to the MKS: internal error
when connecting to the console. The VSphere client and ESXi box are on the same subnet - no firewalls involved. Can telnet to the ESXi IP address on port 902 without issue. It all worked fine for a week, then just stopped.
I did the following
- Restarted the management network on ESXi
- Rebooted the guest
Shutdown the guest via a Terminal Server Session, then tried to power it on via the console. Got another error - A general system error occurred: Unknown error
Finally gave up and rebooted the ESXi box - all then worked OK
The ESXi box is an AD domain member and is resolvable in DNS. It is time synched from the local DC on the same subnet.
This is not a good start when trying to sell VMware on the basis of reliability.
At the moment of the problems did you check any logs on the ESXi host? They will most likely have the reason due this error.
This can either be a DNS issue or a firewall issue but as you mentioned both worked fine. Another problem could be that your session timed out and you received this error.
Did you try disabling the option "config.defaults.security.host.ruissl" ? This might resolve the problem sometimes.
Furthermore do you have enough resources available?
You can find the option under host configuration -> Advanced settings -> Config -> Security -> Host. This MIGHT be the issue (SSL issues).
Furthermore you can find the logs via SSH. SSH into your ESXi host and check the logs under "/var/log".
You should find information in "hostd.log" or "vmkernel.log".
we have exact the same behaviour here at one of our ESXi 5 (768111).
On this box the following operating systems are running:
Windows Server 2008 R2 (as Small Business Server)
Windows 7 Professional
Linux, Fedora Core (Asterisk)
I have the strong suspicion that the problem has something todo with Windows Update being available on the 2008 machine,
because the probleme here occurs always after a few hours (max. few days) after windows update reports updates to be available.
After a reboot everything is fine again until the next month (with new windows updates).
Additionally only the box with SBS is affected, the other ESXi machines (some with Win 7 Professional and Win 7 Ultimate) are fine.
Thanks - no updates received by my guest VM as it gets them from a local WSUS server and haven't approved any this month yet.
Just came back from a few days away and found that the problem had re-occurred. Not only that, but the guest VM itself was no longer accessible from the LAN.
Managed to suspend the VM, put the ESXi box in maintenance mode and rebooted. Console now worked, but the virtual NIC on the guest VM was showing network cable unplugged, yet the management network was working OK, through the same Vswitch and physical NIC.
The physical server is a Dell M605 blade with Broadcom NIC's. I can see from the forum that Broadcom seem to have lots of problems with VMware. I have now disabled Netqueue and the VM is connected to the LAN again. I will have to wait a few days to see if either the console issue or the LAN connectivity problem happen again.
Can't believe I'm having so much trouble with ESXi 5 after over 2 completely trouble free years with 4.
the affected box here is a Lenovo TD230 with Intel Nics (all components and also the TD230 itself without exceptions are on the official VMware HCL).
By the way: the other boxes which are running fine are not built anywhere near to the HCL (Lenovo PCs and Mac Minis under Fusion)
The only special thing with the affected TD230 is that originally there was ESXi 4 installed and then an upgrade to ESXi 5 was installed (but correct and the way which VMware tells us to upgrade).
Under ESXi 4 the problem occured also (but i think the message was a bit different).
I hardly think that is has something todo with the 2008 R2 VM.
Another week or so on and the problem with the console has re-occurred.
Can anyone explain it please, bearing in mind that there are no firewalls involved here and I have the ESX host in DNS ?
I can't allow this to happen if I roll out a full production cluster and back end SAN and at this rate management are going to push for Hyper-V which I think is the wrong decision.
Further investigation has revealed that the problem also affects the esxcli command.
A week or so after a reboot of the ESXi host, when logged in via SSH, the esxcli command will start to return the message
A general system error occurred: Internal error
Only a reboot of the host seems to sort the problem out
I have also redirected the /scratch directory to a larger filesystem in case of space problems, but this has made no difference. The host will work OK for a week or so, then the problem re-occurs.
Help - I really need some assistance with this please.
What storage are you using? I have seen similar behavior during high periods of disk latency - in my case it was NFS on NetApp in a MetroCluster configuration.
I had a same issue, i am able to view VM console from vcenter server, i installed vsphere client on my desktop and i used same vcenter credentials, but i couldn't see VM console, showing same error unable to connect to the MKS: internal error.
finally i found the error is with DNS. Name resolution is not happening. Use nslookup command and check proper DNS connectivity, and also do following changes from this site http://www.gkhan.in/mks-internal-error-on-vmware-console/
The link above is missing a step. I tried it just for fun and the error changed to unable to connect to the MKS: require SSL
The only fix I know for this issue is to restart the management agents. I prefer to do this using SSH:
-but you can do this from the console as well.
1. Power off the VM
2. Go to Resource Allocation and check if the Active memory is only little lesser than the allocated memory
3. If so Increase memory
4. Power on VM.
Hope it helps! Thanks!