Christophe's picture


I try to install several machine on my ESXI4 server and got a problem when running them.

the machine start but dropp to a shell because of some error :

ALERT! /dev/mapper/turnkey-root does not exist. Dropping to a shell

I try to use the vm image and it is the same as the ovf.

Am I missing something.


Jeremy Davis's picture

I wonder if you are experiencing the same issue as these guys? Perhaps try turning off acceleration at boot and see if that stops the issue. Please report back.

Christophe's picture

Disabling acceleration allow to start but there is still some problem.

I am using the Redmine machine, I will try another one

Jeremy Davis's picture

Could you please be more specific about what the problems are after your Redmine appliance boots?

Liraz Siri's picture

Some ESXi users such as Rik Goldman have reported a smooth experience, while others such as yourselves and those in the thread JedMeister linked to have reported a bumpier ride. I'm investigating but it would be useful to have as much information as possible. If you could for example share your hardware configuration and attach screenshots from the problems areas that would be very helpful.
Christophe's picture

First , I forgot to say I am running ESXI4.1 on an HP ML350 G4 dual processor and 12Gb of ram.

I have 2 other machines running for months now without problem on this server (one XP and on ubuntu server 8.10)

With the Turnkey machine :

With acceleration disable, I was able to make the initial setup of the machine (password...) until the last screen where you see the IP address. I was able to change the IP address too.

After that I tried with and without the acceleration without rebooting and have problem:

Trying to connect to the machine from my browser is random.

In webmin I had the display ok one time , I could navigate through the different pages then it display an error on the sql server.

The application (Redmine) give me a Ruby rail error. I never could get it run.

I tried the vm image in vmware workstation and it is ok, everything works fine. So it must be the ovf image which have a problem.

It looks like a lot of problem files not found  from what I saw. I tried differnet type of disk too but it is the same.

I have no issues with ESXi 4.1 when I use the ISO to install, I have not tried OVF for this reason.

Everything fails for me with ESX 4.1 and the only workaround has been to disable acceleration, which is obviously not ideal.

I am happy to volunteer any testing you may want me to do. Just keep in mind I have been trained all my life in M$, so this linux stuff is still pretty new to me. :)

Christophe's picture

Can you try to install a machine (for example Redmine , so we have the same reference) in ovf format through ovf deployement in ESXI4.

There is nothing to know about linux, just setup the machine (answer the question about password) and try to connect to it

I had success deploying the Redmine OVF to ESXi 4.1 (260247).

Exactly what I downloaded:

I just ran through the wizard and left all defaults for the VM and then powered on after copying finished.

Changed passwords and let updates install. I see no issues so far, the redmine page loads and I can log in.

Christophe's picture

I tried again and same problem

Perhaps because of my hardware which is not in the list of supported hardware for ESXI4.1

But I have other machine working for months now..

I will invistigate more.

Just thought I would let you know that my machine is also not listed on the HCL, but I was lucky enough to actually get 4.1 to install without problems.

I have ESXi installed on a Dell OptiPlex GX620.

Jeremy Davis's picture

Be interesting to see if any of them work for you. Would also be useful to confirm that this is actually an upstream Ubuntu bug (rather than specifically a TKL one). This looks like the workaround to try.

Christophe's picture

I upgrade the esxi 4.1 to the latest patch (now build 320137) and it is still the same problem

Christophe's picture

I install a spare server (some model ML350 G4) with esx14.1 plus patch (build 320137), it has only 6Gb of ram (the other one 12).

The HDD are in raid5 (4 disks) the other one was in raid 1 (2 disks) + 1 disk

All other hardware is the same, (dual CPU xeon at 3 Ghz),same raid card HP641, same network card.

And this one is working perfect , using the same ovf I was trying on the othe one. I install 3 machines (redmine, vtiger and revision-control) with no problem at all.

So why ????  I don't know.... (the other one is still working with 2 machine, XP and Ubuntu 8.10)

Jeremy Davis's picture

It does seem strange that 2 very similar machines would react so differently. I would guess that it's something to do with the Ubuntu kernel bug/VMware bug/hardware mix, perhaps its something as simple as different RAM?

Christophe's picture

I am finalizing the install on this server and when everything is working ok, I will shut down the other one and make some more test to try to understand  (hope it will be this week).

I still need to put more ram on this one as I have to move an xp machine from the old server

Jeremy Davis's picture

It'd be great to at least provide a working workaround for this issue if we can't actually fix it (we'll have to wait for an upstream fix I reackon). Also it would be useful if we could find out what the actual offending component is.

Christophe's picture

On the server working, I changed the RAM and install the 12gb of the "non working server", everything went fine.

I put this server in production and it is ok.

On the other server, I install 6GB of ram , change the disk and went to raid5 (4 disks of 72gb), reinstall ESXI4 with all update (now build 348481). I copy the redmine machine from the working server to this one and still the same error. I cannot start the machine on this server. I restart and reset the bios and still not working.

So I don't know what could be wrong as this server was working for 2 years with a winxp and ubuntu 8.10 machine.

L. Arnold's picture

In my experience, though slow, this will get most things to work in VMWare.

You can specify the VM File as the Source, then the ESXI server as the Target.

I am trying a similar task just now (EsXi to EsXi) to see if I can get a resized extended Linux partition, and if nothing else, then I will have a Test Appliance to run Fdisk on and see if I can get a new Partition size.

I have run a few VM conversions w/ converter and it has worked.

More to the immediate task though though, you may also just try to run from ISO Disk Image.  I have been doing this.  Only downfall here is that VMWare tools are not installed by default.  Normally they will take an upgrade from the GUI in VMClient or within SSH/Webmin.

Jeremy Davis's picture

And unfortunately there has still not been an adequite workaround (the only one that seems to work consistently in disabling acceleration during boot).

Going from yarko's troubleshooting it seems that it's a combination of hardware, ESX and Ubuntu (what TKL appliances are based on). I would be inclined to suggest that it's actually a VMware bug.

Add new comment