ESXi 4.1.0 Repair stalls at 25%

[adaptent]

Hi,

I am running ESXi 4.1.0 on a Dell PowerEdge 1850 and had problem after a hard shutdown where the state.tgz file could not be read during startup.

After reviewing the forums it seems the easiest fix is to use the install CD to repair the install.

I booted from the CD several times and each time the repair fails at 25% complete and just hangs there. I have let it go up to 1.5 hours but still the same.

I tried leaving the keyboard both plugged and un-plugged but nada...

Anyone have any suggestions on this?

Thanks in advance.

[DSTAVERT]

How is ESXi installed. Local hard drive, USB? If on HD did you create a separate partition for ESXi or just let it install? You can try installing again to a USB flash stick although if there is still enough left of the old install it may refuse to boot. If it won't boot you would need to alter the old ESXi partitions types but it would get you up again without wiping the old install.

[adaptent]

Hi,

ESXi is installed on the local HD and yes, I just let it install without creating a separate partiton.

What would you suggest for this situation?

Thanks

[DSTAVERT]

Try the install to USB. Use a 1GB or larger disk (larger isn't better since it will only use 1GB). If it installs and boots then you should be good. You can use the install disk to copy off the state.tgz file.

[adaptent]

Hi,

I think I follow and I have the bootable USB stick ready to go now.

Some clarification on the next step though, do you suggest to boot from the USB and then copy the state.tgz file from USB to HD or can I copy the file directly from USB to HD now and then reboot directly from HD?

I am not familiar with the method to copy files in this enviornment.

Thanks again.

[DSTAVERT]

The outcome here is to get you back up and running. I run my environment from USB. I don't install ESXi on hard drives. It may be possible to fix your hard drive install but a mistake can loose the virtual machines you have running on the local hard drives. I am suggesting that you install ESXi directly to the USB disk and run from that. After you are back up and running you can make decisions about what to do next.

Insert the USB stick and begin the install. You may want to disconnect the internal drives?? You will have a choice during the install just choose the USB stick. If ESXi boots and comes back up it will see the original datastore and you can just redo the NICs and re add the Virtual Machines to inventory. If on the reboot ESXi complains about duplicate anything then you will need to edit the partition table and change the partition types on the hard drive and I can help you with that.

[DSTAVERT]

If you need to recover the state.tgz file because you have a complex setup then I would boot with a Linux rescue CD or the ESXi install disk. The linux rescue CD can be one written to another USB stick. It makes it easier to copy the state.tgz file to USB. It can aso be used to modify the ESXi hard drive partition table.

[Dave_Mishchenko]

Have you run a consistency check on the array?  Along the lines of what has been suggested I would boot with a Live CD and see if you can get to state.tgz.  If it's corrupt then ESXi isn't going to be able to do anything it either.

Have you patched or updated the host previously?  If so you could try shift+R during the install boot screen to revert to the prior version.  That way you can at least see if you can get the host to boot and then access the VM data.  

[adaptent]

When I boot it gets to state.tgz and then stops with a PANIC message, saying that it cannot read state.tgz.

From my understanding the easiest way would be to just copy my new state.tgz from the USB stick over the state.tgz from the HD. Then, I should be able to reconfigure the host and re-attach the virtual machines manually.

I just dont know how/where to copy from that one USB location to the HD location in order to attempt this.

[adaptent]

Update:

I wasn't able to boot from the USB because it threw an error about more than one install on the machine.

I used a linux resource CD to search for a state.tgz file that I could replace but wasn't able to locate any such file

It seems the easiest solution is to just rebuld the machine.

I apprecaite the assistance/advice.