ESXi 4/5

[Lackadaisical]

Thick provisioned should give you the best chance of success, I was just wondering if it could be as easy as doing thick vs thin, but of course not. I have a centos system I can try to clone later today when I get home from work, maybe I can figure something out(misplaced optimism...). I'll post back with my results either way. I'm actually quite curious as to why the converter is failing for you. I've done quite a lot of P2V conversions, usually the ones that do not are driver/hardware related (both physical and virtual).

 

Just out of curiosity could you please list the system specs that you converted, and about how long it took to complete the conversion process, and what the "hardware" specs are for the virtual machine (specifically the nic, and HD controller being used).

[greengiant912]

Thick provisioned should give you the best chance of success, I was just wondering if it could be as easy as doing thick vs thin, but of course not. I have a centos system I can try to clone later today when I get home from work, maybe I can figure something out(misplaced optimism...). I'll post back with my results either way. I'm actually quite curious as to why the converter is failing for you. I've done quite a lot of P2V conversions, usually the ones that do not are driver/hardware related (both physical and virtual).

 

Just out of curiosity could you please list the system specs that you converted, and about how long it took to complete the conversion process, and what the "hardware" specs are for the virtual machine (specifically the nic, and HD controller being used).

 

 

Well, I am using an old HP Celeron system that I salvaged to just bring up this tiny server. There is is only one application running on the server, but it was kinda a hassle to setup which is why I was hoping to p2v it, since I already have it the way I want

 

 

The hardware on my ESXi machine is 64bit. The SATA controller works (its an nvidia controller), I was able to create a datastore and everything seems to be transfering fine to the datastore. The motherboard is a ASUS P5N32-E SLI.

 

The nic is an Intel http://www.newegg.com/Product/Product.aspx?Item=N82E16833106123

[Lackadaisical]

What hardware is being presented to the guest OS? You should be able to this information under "Edit Settings" on the VM. I'm just wondering if there is something obvious there that could be the cause of some your issues.

[greengiant912]

What hardware is being presented to the guest OS? You should be able to this information under "Edit Settings" on the VM. I'm just wondering if there is something obvious there that could be the cause of some your issues.

 

 

Memory 1024 (have 8 gigs physical), VMCI device - restricted, SCSI controller0 - LSI logic Parallel, hard disk 1 virtual disk, hard disk 2 virtual disk, cd/dvd drive, network adapter 1 - vm network.

 

 

omg I am wondering if the converter took my swap partition and put it on it own hard drive...

[Lackadaisical]

Interesting ...having more than one drive in the system that had to be converted could also be part of the issue. It might be worth launching the converter again and looking through all the options again just to see if you can find out if it was putting the partitions in the right places ( I think it did). A more likely cause would be your system not liking the hardware that is being presented to it.

 

 

There are a couple of options I would say you should try, and it depends how much you have already messed with the current P2V clone (if you have done a lot of things to it, it might have been corrupted). I would suggest that you copy just the vmdk's, to a new location. Then create a completely new VM, using the Custom wizard. Keep the hardware as simple/accurate as possible. For the OS just choose other for the time being, set the NIC to be the E1000 adapter, use the LSI Logic Parallel disk controller, and then you will select the option to use existing disks. You will have to manually add the second disk after the wizard, either by editing settings again (or check the box that lets you edit after creation). Do not power it back on until you have the second disk added.

[greengiant912]

Interesting ...having more than one drive in the system that had to be converted could also be part of the issue. It might be worth launching the converter again and looking through all the options again just to see if you can find out if it was putting the partitions in the right places ( I think it did). A more likely cause would be your system not liking the hardware that is being presented to it.

 

 

There are a couple of options I would say you should try, and it depends how much you have already messed with the current P2V clone (if you have done a lot of things to it, it might have been corrupted). I would suggest that you copy just the vmdk's, to a new location. Then create a completely new VM, using the Custom wizard. Keep the hardware as simple/accurate as possible. For the OS just choose other for the time being, set the NIC to be the E1000 adapter, use the LSI Logic Parallel disk controller, and then you will select the option to use existing disks. You will have to manually add the second disk after the wizard, either by editing settings again (or check the box that lets you edit after creation). Do not power it back on until you have the second disk added.

 

Here are the options currently... Why is the boot partition on a separate disk from the swap, and data partitions?

Edited September 12, 2012 by greengiant912

[Lackadaisical]

That actually looks correct, that's what I would have expected. It's just the way that Linux is handling the HD partitioning. You have your boot, and everything else is thrown on to the LVM, that was made when you installed the OS. I wonder if the option to "reconfigure destination" could have caused an issue. If it interpreted things wrong and configured it for the wrong type of system this might be an issue. When you initially go through the converter, does it properly identify the physical machine information?

 

I'll run through my converter, and see if I can find anything that might be useful too...

[greengiant912]

That actually looks correct, that's what I would have expected. It's just the way that Linux is handling the HD partitioning. You have your boot, and everything else is thrown on to the LVM, that was made when you installed the OS. I wonder if the option to "reconfigure destination" could have caused an issue. If it interpreted things wrong and configured it for the wrong type of system this might be an issue. When you initially go through the converter, does it properly identify the physical machine information?

 

I'll run through my converter, and see if I can find anything that might be useful too...

 

Are you running 6.2?

 

 

I am also going to try building another machine from scratch to see if I have issues even running something.

[Lackadaisical]

I have 5.3 and 6.3, unfortunately nothing with 6.2, and both systems are x64.

[Lackadaisical]

My luck seems to be even worse than yours. For me the converter is not even detecting my LVG properly and all I end up with is it seeing my /boot. Not very useful.... At least yours does see all of the expected partitions so that's a good place to be (comparatively speaking). Since it looks like it is capturing all of the partitions correctly for you, I still want to lean towards it being a "hardware" related issue. Most likely the VM does not like what is being presented to it.

 

My older 5.3 centos system doesn't actually even boot at this point either, so I don't know if my issue with the converter is just being caused by having the newer version of centos, or something else entirely...

 

Edit: what happens if you uncheck the "Create optimized partition layout" ?

Edited September 12, 2012 by Lackadaisical

[greengiant912]

[Lackadaisical]

In case anyone actually runs into this problem (which I doubt), I'll put this here for future reference. The main point is that VMware converter 5 does not support Logical Volumes on CentOS.

 

http://communities.vmware.com/message/1816684#1816684

 

You can get around this in a round about way by following the first option in the above post (which only works on some versions of CentOS; worked on 6.2 but not 6.3). If that fails the other option is to rebuild the initrd image manually using option #2.

 

In the end you may have a working(booting) VM. At that point you will also find yourself without a working network interface, as both greengiant and I found out... In the end what happened was the VM trashed eth0 and marked it as being an inactive device. Eth1 became the new active adapter, you must manually create ifcfg-eth1 file, with the correct network information and then restart the network service.

 

This is definitely not to encourage anyone into actually doing this, but if you are really desperate and like to bang your head against walls for long periods of time; you might just make it out at the end with a P2V CentOS system that actually works, maybe...