VMware Cloud Community
billymcc
Contributor
Contributor
Jump to solution

Unable to obtain hardware information for selected machine (Ubuntu 64 phsyical)

Hello,

I need to p2v a Ubuntu 10.10, 64 bit physical and I am getting the ever so common "Unable to obtain hardware information for selected machine" error.

I have attached the log files I also noticed that in the auth.log of the server when you try to do the conversiton the following lines stream by:

Dec  1 11:13:42 seg-dev sshd[9830]: Accepted password for root from x.xx.xx.xx port 63552 ssh2
Dec  1 11:13:42 seg-dev sshd[9830]: pam_unix(sshd:session): session opened for user root by (uid=0)
Dec  1 11:13:42 seg-dev sshd[9830]: pam_unix(sshd:session): session closed for user root
Dec  1 11:14:01 seg-dev CRON[9922]: pam_unix(cron:session): session opened for user root by (uid=0)
Dec  1 11:14:01 seg-dev CRON[9922]: pam_unix(cron:session): session closed for user root

I dont see any failures to autenticate, and the machine the converstion client/server is running on can ssh to the machine to be converted and log on as root with out issue.

I get the same error if I disable the linux firewall service running on the machine (ufw).

All in all this is a fairly simple vanilla LAMP server but it wont seem to convert. Please help!

Billy

0 Kudos
1 Solution

Accepted Solutions
ivivanov
Expert
Expert
Jump to solution

Hi Billy,

I think I found the problem - it is in your /storage volume. If you look at the sysinfo log file it says

-->   <volume mountPoint="/storage" blockSizeInBytes="4096" capacityInBlocks="105819486" blocksFree="47303375" blocksUsed="58516111" devicePath="/dev/cciss/c0d1" fileSystemType="ext4" isBootVolume="false" label=""/>

If you look at the device path it is devicePath="/dev/cciss/c0d1". Later looking at the disks section in sysinfo output there is /dev/cciss/c0d1 disk which has /dev/cciss/c0d1p1 partition. It means the whole *disk* is used as a volume, but not only the partition. Unfortunately Converter does not support such configurations. What happens is it tries to populate hardwareInfo.storage.volume.diskId field for each volume and it assumes each volume is located either within a partition on a disk or within a LVM logical volume. Since for the /storage volume none of these is true (it is neither within a partition nor within a logical volume) the hardwareInfo.storage.volume.diskId field is left unset and this is the error that eventually is reported:

2011-12-02T10:35:20.989-05:00 [198536 error 'Default'] Not initialized: converter.info.ComputerHardwareInfo hardwareInfo.storage.volume.diskId

What I can think of is to run the conversion again and try to deselect the /storage volume, convert the operating system and after (if) the conversion completes successfully, you can boot the target, re-create the /storage volume manually and then copy the data from the source.

Hope this helps,

Ivan

__________
It is worse!

View solution in original post

0 Kudos
22 Replies
ivivanov
Expert
Expert
Jump to solution

That is strange. Looking at the server log there is a strange entry there

2011-11-30T09:31:08.388-05:00 [118264 error 'vmomi.soapStub[4]']
--> Required property diskId is missing from data object of type ConverterStorageInfoVolumeInfo

It means a data structure has not been filled properly. Unfortunately there is no Worker log in the archive, and it is the interesting part. Can you get it manually - it should be on the Converter Server machine under %ALLUSERSPROFILE%\Application Data\VMware\VMware vCenter Converter Standalone\logs (or under ProgramData on Vista+).

Another thing you could try for troubleshooting is to run manually vmware-sysinfo executable on the source. there is a vmware-sysinfo-lin64.tar.gz file located in the Converter installation directory under %PROGRAMFILES(X86)%\VMware\VMware vCenter Converter Standalone directory. You can copy it to the source machine, extract it to a temp folder and run it from there (there is a wrapper shell script - you should start the script and it sets the LD_LIBRARY_PATH and whatever else is needed). This executable detects the source information in terms of hardware (disks, networks, CPUs, memory) and software (distribution, kernel, volumes). It si automatically deployed and ran by Converter Server on the source through SSH and as a result generates an XML describing the source, which is later parsed by Converter (in your case this step is failing). It also generates a log file in the same directory. You can post this file here either.

Based on worker log and sysinfo log we can investigate in more details what caused source query to fail.

__________
It is worse!
0 Kudos
ranjitcool
Hot Shot
Hot Shot
Jump to solution

Alright there might be a few issues.

My experience told me that vmware ONLY supports rhel P2V.

Now with that out of our way, I would recommend doing a cold clone for this ubuntu box.

How ever looking at your logs a few things came up.

[198612 error 'Default'] Found dangling SSL error: [0] error:00000001:lib(0):func(0):reason(1)  -- There is a name resolution issue on your physical server, check /etc/hosts (going by my rhel knowledge, dont expect ubuntu to place it else where)  [kb here]

[198264 error 'vmomi.soapStub[1]']
--> Required property diskId is missing from data object of type ConverterStorageInfoVolumeInfo  -- Now this is reported when vss cannot create a snapshot of the vm. Now this can be fixed by having vmware tools installed. (I never really tried it but I don't see why not). You will have trouble running vss if your physical box has no vmtools installed as it cannot quience the disk. [kb here]

Like I said, best bet is to use cold clone disk - remember to slipstream scsi controller drivers and also do the post process so the disk paths are changed after the conversion is done.

Thanks

RJ

Please award points if you find my answers helpful Thanks RJ Visit www.rjapproves.com
billymcc
Contributor
Contributor
Jump to solution

Here is the output of running that script on the server i need to p2v.

0 Kudos
ivivanov
Expert
Expert
Jump to solution

Can you also attach the vmware-sysinfo.log file that should be generated in the same directory as well as the Worker log file?

__________
It is worse!
0 Kudos
billymcc
Contributor
Contributor
Jump to solution

Yep! thanks so much for your help!

0 Kudos
billymcc
Contributor
Contributor
Jump to solution

By the way, i checked and the host name of the server is already in the /etc/hosts file, but i went ahead and defined the entry for the converter machine too, and I still get the same error...

0 Kudos
billymcc
Contributor
Contributor
Jump to solution

Cold Clone? is that the converter-helper-vm-x64.iso image that is in the directory?

I googled cold clone, and saw that earlier versions of the converter had it with it.

DId they change the name?

Billy

0 Kudos
ivivanov
Expert
Expert
Jump to solution

I am still missing the Worker log file? From sysinfo log it looks like all devices are recognized properly, there is only a problem with grub device map (because Ubuntu 10.10 is with GRUB2 I guess).

__________
It is worse!
0 Kudos
ivivanov
Expert
Expert
Jump to solution

Nope, converter-helper-vm-x64.iso is mounted as a CD-ROM image on the target when performing Linux P2V and the target is reconfigured to boot from CD. There is a Converter Helper service in the ISO file that performs the actual cloning.

Cold-clone is no longer supported and it has never been relased with standalone versions of Converter.

__________
It is worse!
0 Kudos
billymcc
Contributor
Contributor
Jump to solution

I will dig for that file today I rebooted the box, yesterday and now it hangs at the grub screen?

I didnt make any mods so I dont know what is going on.


Billy

0 Kudos
billymcc
Contributor
Contributor
Jump to solution

Where is the worker log location?

I re-ran it this morning and no worker log gets created?

-Billy

0 Kudos
patanassov
VMware Employee
VMware Employee
Jump to solution

It's in %ALLUSERSPROFILE%\Application Data\VMware\VMware vCenter Converter Standalone\logs, together with the server's log.

However it is better to export the logs from the failed task (not job), the bundle will contain also the helper's log.

0 Kudos
billymcc
Contributor
Contributor
Jump to solution

ivivanov had me run it from the system mannually because it was failing, I looked in the exported bundle and the worker log is not there.

0 Kudos
ivivanov
Expert
Expert
Jump to solution

It should be under %ALLUSERSPROFILE%\Application Data\VMware\VMware vCenter Converter Standalone\logs on the machine where Converter Server is installed. On Vista and above the location should be under ProgramData\VMware\...

__________
It is worse!
0 Kudos
billymcc
Contributor
Contributor
Jump to solution

No, "~re\VMware vCenter Converter Standalone\logs " exists...

See screenshot attached.

0 Kudos
ivivanov
Expert
Expert
Jump to solution

Not in  "Program Files" but "ProgramData" folder...

__________
It is worse!
0 Kudos
billymcc
Contributor
Contributor
Jump to solution

oops, sorry i see it now...

Here is the latest of both log bundles

0 Kudos
ivivanov
Expert
Expert
Jump to solution

Hi Billy,

I think I found the problem - it is in your /storage volume. If you look at the sysinfo log file it says

-->   <volume mountPoint="/storage" blockSizeInBytes="4096" capacityInBlocks="105819486" blocksFree="47303375" blocksUsed="58516111" devicePath="/dev/cciss/c0d1" fileSystemType="ext4" isBootVolume="false" label=""/>

If you look at the device path it is devicePath="/dev/cciss/c0d1". Later looking at the disks section in sysinfo output there is /dev/cciss/c0d1 disk which has /dev/cciss/c0d1p1 partition. It means the whole *disk* is used as a volume, but not only the partition. Unfortunately Converter does not support such configurations. What happens is it tries to populate hardwareInfo.storage.volume.diskId field for each volume and it assumes each volume is located either within a partition on a disk or within a LVM logical volume. Since for the /storage volume none of these is true (it is neither within a partition nor within a logical volume) the hardwareInfo.storage.volume.diskId field is left unset and this is the error that eventually is reported:

2011-12-02T10:35:20.989-05:00 [198536 error 'Default'] Not initialized: converter.info.ComputerHardwareInfo hardwareInfo.storage.volume.diskId

What I can think of is to run the conversion again and try to deselect the /storage volume, convert the operating system and after (if) the conversion completes successfully, you can boot the target, re-create the /storage volume manually and then copy the data from the source.

Hope this helps,

Ivan

__________
It is worse!
0 Kudos
billymcc
Contributor
Contributor
Jump to solution

This makes complete since...

How do I "deslect" the /storage ?

It would be easy enough to make a disk for this after I boot the target vm. and just copy over the contents of /storage on the source... i can see how that could easly work.

I just dont know how to deselect that with the converter tool?

Billy

0 Kudos