[chef] Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Problem running chef-client, just from 2 days ago.

[Carlos Camacho < >]

Can you send the client.rb of your node?

On Wed, Jul 30, 2014 at 3:28 PM, Kalle Lehtonen 
<
 >
 wrote:
> Hi,
>
> Just ran on performance server with "chef-client -S <chef-server name>
> -N <node name> and I still get that EOFError.
>
> We tested as well with standard flavors and those were successful as
> you mentioned in your e-mail to the list on 9th of July.
>
> Cheers,
> Kalle
>
> 2014-07-30 13:02 GMT+03:00 Carlos Camacho 
> <
 >:
>> Hello,
>>
>> Again, the problem its that you should not have the node_name
>> parameter in your client.rb file.
>> Do it, and you will fix the problem in your Performance images.
>> Its not about the Rackspace images as they replace the images to get
>> the latest updates once in a while.
>>
>>
>> Cheers.
>>
>>
>>
>>
>>
>>
>> On Wed, Jul 30, 2014 at 10:31 AM, Kalle Lehtonen
>> <
 >
>>  wrote:
>>> Hi Carlos,
>>>
>>> Sorry to comeback to this issue this late.
>>>
>>> But would you be able elaborate more on the issue?
>>>
>>> We have encountered the same issue (and still encounter) in Rackspace
>>> environment with Windows 2008R2 images with Rackconnect enabled servers. 
>>> The
>>> only difference was that we were able to try Windows client with 
>>> non-Rackspace
>>> server which worked. Also Chef client and server both running CentOS 6.5 
>>> PVHVM
>>> image worked correctly.
>>>
>>>
>>> I cannot find either of those server IDs on the list I get from Rackspace.
>>> Could you tell the names of those images?
>>> You also mentioned that the server flavor had an impact to it. I would 
>>> then
>>> guess you used standard flavors? Was it required for both client and 
>>> server?
>>>
>>> It would be also interesting to know if Rackspace was planning to fix the
>>> performance flavor images and was this only affecting clients, server or 
>>> both?
>>>
>>> I am just interested on this because I would hate to go back to standard
>>> servers in Rackspace.
>>>
>>> Cheers,
>>> Kalle
>>>
>>> 2014-07-30 9:39 GMT+03:00 Carlos Camacho 
>>> <
 >:
>>>> Hello Kalle,
>>>>
>>>> No problem man, this was a really nasty issue with a really simple fix 
>>>> (Really).
>>>> Them problems its that ohai some times get crazy getting the node name
>>>> for Windows Performance images on Rackspace.
>>>> This is the fix,
>>>>
>>>> You can either bootstrap your chef-client with the -N shortcut (My 
>>>> choice):
>>>>
>>>> knife bootstrap windows winrm IP -x Administrator -P pass -N hostname
>>>>
>>>> Or add to client.rb:
>>>>
>>>> node_name "hostname"
>>>>
>>>>
>>>> I hope that this helps,
>>>>
>>>> Carlos.
>>>>
>>>>
>>>>
>>>>
>>>> On Tue, Jul 29, 2014 at 5:23 PM,  
>>>> <
 >
>>>>  wrote:
>>>>> Hi Carlos,
>>>>>
>>>>> Sorry to comeback to this issue this late.
>>>>>
>>>>> But would you be able elaborate more on the issue?
>>>>>
>>>>> We have encountered the same issue (and still encounter) in Rackspace
>>>>> environment with Windows 2008R2 images with Rackconnect enabled 
>>>>> servers. The
>>>>> only difference was that we were able to try Windows client with 
>>>>> non-Rackspace
>>>>> server which worked. Also Chef client and server both running CentOS 
>>>>> 6.5 PVHVM
>>>>> image worked correctly.
>>>>>
>>>>>
>>>>> I cannot find either of those server IDs on the list I get from 
>>>>> Rackspace.
>>>>> Could you tell the names of those images?
>>>>> You also mentioned that the server flavor had an impact to it. I would 
>>>>> then
>>>>> guess you used standard flavors? Was it required for both client and 
>>>>> server?
>>>>>
>>>>> It would be also interesting to know if Rackspace was planning to fix 
>>>>> the
>>>>> performance flavor images and was this only affecting clients, server 
>>>>> or both?
>>>>>
>>>>> I am just interested on this because I would hate to go back to standard
>>>>> servers in Rackspace.
>>>>>
>>>>> Cheers,
>>>>> Kalle