[rudder-dev] new autofs techniques

Nicolas Charles nicolas.charles at normation.com
Thu Apr 11 17:06:20 CEST 2013


Oh, I realize I never answered to this mail

On 26/03/2013 13:10, Michael Gliwinski wrote:
> On Tuesday 26 Mar 2013 12:07:41 Nicolas Charles wrote:
>> On 26/03/2013 10:54, Michael Gliwinski wrote:
>>> I read the comments on github, I must admit, I'm not really sure what
>>> TRACKINGVARIABLE and TRACKINGKEY are.  However that does remind me that
>> They are the most complex part of the technique writing. When the
>> promises are generated, Rudder "aggregates" together all the variables
>> of all the directives based on the same Techniques; and it flattens all
>> variables in multi-values sections.
>> So we need to keep a track of which variable value comes from with
>> Directives, and there come the TRACKINGVARIABLE, which have the size of
>> another variable if it is multivalued
> OK, that makes sense, thanks.  Just one question, what do you mean by "size"
> in this context?  (is it maybe the number of values coming from the same
> particular directive?)

Yes, the size is the number of values

>> With a small example issued from checkGenericFileContent, we have the
>> TRACKINGVARIABLE with a SAMESIZEAS GENERIC_FILE_CONTENT_PATH
>> GENERIC_FILE_CONTENT_PATH is in a mutlivalued sections, meaning we can
>> have several for each Directives
>>
>> If we have three directives, configured as follow :
>>
>> DirectiveA : file 1, file2
>> DirectiveB : file 3
>> DirectiveC : file 4, file 5, file 6
>>
>> GENERIC_FILE_CONTENT_PATH will end up being a list of 6 elements :
>> file 1, file2, file 3, file 4, file 5, file 6
>> And TRACKINGKEY will also be a list of 6 elements :
>> DirectiveA, DirectiveA, DirectiveB, DirectiveC, DirectiveC, DirectiveC
>>
>> So that we can iterate over each elements, as well as the reporting, to
>> get proper report for each file
> Thanks very much for explaining, I understand better now :)

Great, I'm glad it helped

Nicolas


More information about the rudder-dev mailing list