You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@vcl.apache.org by David DeMizio <dd...@ncf.edu> on 2015/05/04 16:02:53 UTC
Re: capture image
Hi Andy,
Thanks, that got me a bit further. Now it fails at some other part. The
clone is still running but VCL must of timed out or something. I see a few
SOAP errors and read errors. I have attached the vcld log
On Tue, Apr 28, 2015 at 6:26 PM, Andy Kurth <an...@ncsu.edu> wrote:
> The SDK problems can probably be fixed by downgrading the libwww-perl
> (LWP) Perl module installed on your management node. First, check the
> version you have installed:
> *perl -e 'use LWP; print "$LWP::VERSION\n";'*
>
> (This is probably going to display 6.<something>)
>
> Then run the following:
> *cpan GAAS/libwww-perl-5.837.tar.gz*
>
> Run the perl -e command again. It hopefully reports 5.837. Then try to
> restart the image capture. From the vcld.log files you sent, it looks like
> you may be running vcld -setup each time. You don't need to do this as it
> creates multiple useless image and other entries in the database. All you
> need to do is change the request.stateid value back to 16 (image). For the
> last log you sent the request ID is 10 so run the following on your
> database server:
> *mysql -e 'UPDATE request SET stateid = 16 WHERE id = 10;' vcl*
>
> Immediately after running this, start up *tail -f /var/log/vcld.log*. It
> should attempt to capture the image again.
>
> BTW - VCL 2.4.2 includes a new feature which allows you to restart a
> failed image capture from the website.
>
> Regards,
> Andy
>
> On Tue, Apr 28, 2015 at 1:23 PM, David DeMizio <dd...@ncf.edu> wrote:
>
>> Hi Andy,
>>
>> I have attached some files. I tried the Vsphere SDK 5.5 and 5.5u2 as
>> well. both give same error. Soap request error as you will see in the logs
>> along with uncaught perl errors in VMware.pm. was there an update to
>> VMware.pm? to handle 5.5?
>>
>>
>>
>>
>>
>> On Tue, Apr 28, 2015 at 10:25 AM, Andy Kurth <an...@ncsu.edu> wrote:
>>
>>> Hi David,
>>>
>>> On Tue, Apr 28, 2015 at 8:45 AM, David DeMizio <dd...@ncf.edu> wrote:
>>>>
>>>> Does the vm I'm trying to capture have to be in /vmfs/volumes/VCL-WORKSPACE?
>>>> I don't recall it having to be there as the VMware module looks at all the
>>>> shared datastores but I could be wrong. Thanks
>>>>
>>>
>>> You are correct. It should not matter which datastore contains the VM
>>> you're capturing for a base image. I took a closer look at the log file.
>>> I believe the underlying problem is related to these lines:
>>>
>>> 2015-04-20
>>>> 15:27:55|4268|41:41|image|vSphere_SDK.pm:get_registered_vms(211)|found 0
>>>> registered VMs:
>>>> 2015-04-20
>>>> 15:27:55|4268|41:41|image|VMware.pm:get_vmx_file_paths(5247)|found 0
>>>> registered vmx files
>>>
>>>
>>> We need to figure out why the *get_registered_vms* subroutine is
>>> failing to find any VMs. It does not care which datastore is used but does
>>> care which datacenter and resource pool the VM resides in. In your log
>>> file, it is looking for VMs under the "Computing Lab" resource pool which
>>> is under the datacenter named "New College":
>>>
>>> 2015-04-20
>>>> 15:27:46|4268|41:41|image|vSphere_SDK.pm:_get_resource_pool_view(2798)|single
>>>> resource pool found on VM host vdi-vcenter will be used: /New
>>>> College/Computing Lab
>>>> 2015-04-20
>>>> 15:27:47|4268|41:41|image|vSphere_SDK.pm:_get_datacenter_view(2628)|found
>>>> datacenter VM host on vdi-vcenter: New College
>>>
>>>
>>> Can you confirm that the VM is indeed in this location according to the
>>> "Hosts and Clusters" view? It may be helpful if you could send a
>>> screenshot of the "VMs and Templates" and "Hosts and Clusters" views from
>>> your vSphere Client.
>>>
>>> The mismatched SDK version may also be the problem. I'd recommend
>>> attempting to get the 5.5 SDK installed. I know recent versions can be
>>> finicky since the installer has trouble installing some Perl modules and
>>> some Perl modules changed their SSL requirements. First remove the 5.0 SDK
>>> by running vmware-uninstall-vSphere-CLI.pl. Also remove any other VMware
>>> CLI/API/SDK products which may be installed that appear when you type
>>> vmware-uninstall <tab><tab>. Try to install the 5.5 SDK. Below are some
>>> notes I had taken when encountering SDK installation issues. Try running
>>> these commands:
>>>
>>> yum install -y perl-CPAN perl-Archive-Zip perl-Class-MethodMaker
>>> perl-Data-Dump uuid-perl libuuid-devel glibc.i686 openssl-devel
>>> export http_proxy=
>>> export ftp_proxy=
>>> export PERL_LWP_SSL_VERIFY_HOSTNAME=0
>>> ./vmware-install.pl -d
>>>
>>> Hope this helps,
>>> Andy
>>>
>>
>>
>
Re: capture image
Posted by David DeMizio <dd...@ncf.edu>.
Looks like upgrading Net-HTTP to 6.03 fixed the issue
On Mon, May 4, 2015 at 10:02 AM, David DeMizio <dd...@ncf.edu> wrote:
> Hi Andy,
>
> Thanks, that got me a bit further. Now it fails at some other part. The
> clone is still running but VCL must of timed out or something. I see a few
> SOAP errors and read errors. I have attached the vcld log
>
>
>
>
> On Tue, Apr 28, 2015 at 6:26 PM, Andy Kurth <an...@ncsu.edu> wrote:
>
>> The SDK problems can probably be fixed by downgrading the libwww-perl
>> (LWP) Perl module installed on your management node. First, check the
>> version you have installed:
>> *perl -e 'use LWP; print "$LWP::VERSION\n";'*
>>
>> (This is probably going to display 6.<something>)
>>
>> Then run the following:
>> *cpan GAAS/libwww-perl-5.837.tar.gz*
>>
>> Run the perl -e command again. It hopefully reports 5.837. Then try to
>> restart the image capture. From the vcld.log files you sent, it looks like
>> you may be running vcld -setup each time. You don't need to do this as it
>> creates multiple useless image and other entries in the database. All you
>> need to do is change the request.stateid value back to 16 (image). For the
>> last log you sent the request ID is 10 so run the following on your
>> database server:
>> *mysql -e 'UPDATE request SET stateid = 16 WHERE id = 10;' vcl*
>>
>> Immediately after running this, start up *tail -f /var/log/vcld.log*.
>> It should attempt to capture the image again.
>>
>> BTW - VCL 2.4.2 includes a new feature which allows you to restart a
>> failed image capture from the website.
>>
>> Regards,
>> Andy
>>
>> On Tue, Apr 28, 2015 at 1:23 PM, David DeMizio <dd...@ncf.edu> wrote:
>>
>>> Hi Andy,
>>>
>>> I have attached some files. I tried the Vsphere SDK 5.5 and 5.5u2 as
>>> well. both give same error. Soap request error as you will see in the logs
>>> along with uncaught perl errors in VMware.pm. was there an update to
>>> VMware.pm? to handle 5.5?
>>>
>>>
>>>
>>>
>>>
>>> On Tue, Apr 28, 2015 at 10:25 AM, Andy Kurth <an...@ncsu.edu>
>>> wrote:
>>>
>>>> Hi David,
>>>>
>>>> On Tue, Apr 28, 2015 at 8:45 AM, David DeMizio <dd...@ncf.edu>
>>>> wrote:
>>>>>
>>>>> Does the vm I'm trying to capture have to be in /vmfs/volumes/VCL-WORKSPACE?
>>>>> I don't recall it having to be there as the VMware module looks at all the
>>>>> shared datastores but I could be wrong. Thanks
>>>>>
>>>>
>>>> You are correct. It should not matter which datastore contains the VM
>>>> you're capturing for a base image. I took a closer look at the log file.
>>>> I believe the underlying problem is related to these lines:
>>>>
>>>> 2015-04-20
>>>>> 15:27:55|4268|41:41|image|vSphere_SDK.pm:get_registered_vms(211)|found 0
>>>>> registered VMs:
>>>>> 2015-04-20
>>>>> 15:27:55|4268|41:41|image|VMware.pm:get_vmx_file_paths(5247)|found 0
>>>>> registered vmx files
>>>>
>>>>
>>>> We need to figure out why the *get_registered_vms* subroutine is
>>>> failing to find any VMs. It does not care which datastore is used but does
>>>> care which datacenter and resource pool the VM resides in. In your log
>>>> file, it is looking for VMs under the "Computing Lab" resource pool which
>>>> is under the datacenter named "New College":
>>>>
>>>> 2015-04-20
>>>>> 15:27:46|4268|41:41|image|vSphere_SDK.pm:_get_resource_pool_view(2798)|single
>>>>> resource pool found on VM host vdi-vcenter will be used: /New
>>>>> College/Computing Lab
>>>>> 2015-04-20
>>>>> 15:27:47|4268|41:41|image|vSphere_SDK.pm:_get_datacenter_view(2628)|found
>>>>> datacenter VM host on vdi-vcenter: New College
>>>>
>>>>
>>>> Can you confirm that the VM is indeed in this location according to the
>>>> "Hosts and Clusters" view? It may be helpful if you could send a
>>>> screenshot of the "VMs and Templates" and "Hosts and Clusters" views from
>>>> your vSphere Client.
>>>>
>>>> The mismatched SDK version may also be the problem. I'd recommend
>>>> attempting to get the 5.5 SDK installed. I know recent versions can be
>>>> finicky since the installer has trouble installing some Perl modules and
>>>> some Perl modules changed their SSL requirements. First remove the 5.0 SDK
>>>> by running vmware-uninstall-vSphere-CLI.pl. Also remove any other VMware
>>>> CLI/API/SDK products which may be installed that appear when you type
>>>> vmware-uninstall <tab><tab>. Try to install the 5.5 SDK. Below are some
>>>> notes I had taken when encountering SDK installation issues. Try running
>>>> these commands:
>>>>
>>>> yum install -y perl-CPAN perl-Archive-Zip perl-Class-MethodMaker
>>>> perl-Data-Dump uuid-perl libuuid-devel glibc.i686 openssl-devel
>>>> export http_proxy=
>>>> export ftp_proxy=
>>>> export PERL_LWP_SSL_VERIFY_HOSTNAME=0
>>>> ./vmware-install.pl -d
>>>>
>>>> Hope this helps,
>>>> Andy
>>>>
>>>
>>>
>>
>