datastore suddenly goes inactive. VM Error - Unable to connect to MKS:

[Raman_yag]

Hi,

I am getting this problem multiple times now. And every time i had to reinstall the ESXi and had to redo the VM's from scratch.

I have a HP proliant g7 server with 4 300GB drives with RAID and visible as single 1.09TB drive.

I had previously installed 4.0ESXi and had 4 RHEL6.1 VM's runnings without any problems for almost 6 monts.

Suddunly one day i was unable to access the VM's. When i try to access the console from vSphere Client I get the error.

Unable to connect to MKS: Virtual machine config file does not exist.

The VM's could not be stopped anymore.

When i try to look into the datastore i would see it empty.

Rescanning the datastore results in datastore being inactive.

I reinstalled the ESXi and tried to import the VM's that i had exported earlier.

The VM would not start after deploying the template. I am able to deploy the same VM on another (Dell) server. This is the VM that was originally exported from the same HP VM that went bad.

So i had to re-create the VM from scratch, but after just a few days the same problem re-surfaces.

This time the installed the ESXi 5.0 and during installation chose the option to retain the existing datastore.

I was able to retain the VM's and dint have to recreate them from scratch. I was happy thinking i have gotten over the issue and that the problem must have been with the 4.0ESXi.

But after 2 days the same problem re-surfaced.

Now the only easy fix that i see is to re-install the ESXi again and hope to retain the VM's.

I have tried the solution as explained in this post by restarting the management agent but that does not work.

http://www.virtualization.net/1513-unable-to-connect-to-mks-virtual-machine-config-file-does-not-exi...

Please let me know if anyone knows what to do here.

Is it possible there is problem with the physical disks on this machine?

Raman

[GreatWhiteTec]

Do you have any firewalls in place. MKS connectivity it usually ends up being port 903 for ESX(i) not being open or port 902 for ESXi 5.

[GreatWhiteTec]

If firewall is not the issue. You can try to recreate your vmx file.

[Raman_yag]

No firewall is not the issue, and there is no way to recreate the vmx file as the datastore is inaccessible.

Its as if the datastore is lost.

Also here is something else that i found which may indicate its a hardware problem with the HP disk array.

When i rebooted the system and looked at the ROM messages, it indicated that the logical drive has been lost or corrupted.

It then proceeded to offer to press F1 to continue or F2 to re-instate the logical drive (it may destroy the contents).

I have to press F2 to proceed. I re-installed the ESXi and had selected to retain the existing datastore.

[chriswahl]

Sounds like your disk array is suffering some issues, check the logs to find out why the array has failed.

I have a HP proliant g7 server with 4 300GB drives with RAID and visible as single 1.09TB drive.

Based on the fact that you had 4 x 300GB drives with 1.09 TB usable, I'm guessing you're not using a RAID that has a parity drive. As (4 - 1) x 300 = < 900 GB of space. It would behoove you to use a parity (or mirror) based RAID array (such as RAID 5, 6 or 1+0).

The MKS (mouse, keyboard, screen) error is just stating that it could not find the VM to connect to it for console access. It's a symptom of the root cause, not an issue in itself.

[vGuy]

you may also want to ensure all your drives and BBWC are in healthy state..