The PlatformLabeler plugin has been a real help for some of those cases of
OS and platform values, and it handles them automatically.


On Tue, Apr 29, 2014 at 9:42 AM, Jeff <[email protected]> wrote:

> Maybe this exists but I missed it, but I also wonder if there is a way to
> inject system environment variables on any node (master/slaves) so that
> things like OS name/family, architecture, OS distribution (e.g., 'Ubuntu
> 14.04 LTS'), etc. could be already available.
>
> Seems redundant to have to manually specify labels that are inherent to
> the slave environment.
>
>
> On Tue, Apr 29, 2014 at 7:40 AM, Kevin Fleming (BLOOMBERG/ 731 LEXIN) <
> [email protected]> wrote:
>
>> You are correct, the decision logic for choosing slaves only takes labels
>> into account, it does not match up JDK requirements. You'll have to put the
>> appropriate labels on your slaves as well.
>>
>> Rather than modifying the decision logic, though, a way to solve this
>> problem would be for the configuration code that manages the list of
>> installed JDKs on a slave to have an option to inject labels automatically
>> into the slave's list of labels based on the installed JDKs. That would
>> allow the normal label-based decision logic to work, but would eliminate
>> the duplicate data entry (and the opportunity for mistakes).
>>
>>
>> ----- Original Message -----
>> From: [email protected]
>> To: [email protected]
>> At: Apr 29 2014 07:51:24
>>
>> As we need a lot of manually installed tools, we decided to have strictly
>> separated slaves and provide rather lots of labels to get each project
>> built on the right slave.
>> This works really good so far. Now we thought we could spare at least the
>> label for the JDK, as Jenkins in a sense has all needed information to pick
>> the right slave for a needed JDK:
>> * All used JDKs are provided in the Jenkins core configuration, and all
>> of them are marked as "do not install automatically".
>> * Each project has the "Used JDK" set to a particular JDK.
>> * Each node has a list of all manually installed JDKs.
>> Hence, if Jenkins needs to pick a Slave from the list of possible nodes,
>> it can "see" which of those have the needed JDK for a particular project
>> installed.
>> Unfortunately it does not. In fact, it picks just ANY, even one that has
>> not the needed JDK installed.
>>
>> Certainly this can be easily worked around by using another label for the
>> JDK, but we think this is redenundant information. If Jenkins can
>> auto-install a missing JDK, why can't it in turn not simply use a different
>> slave if the JDK is missing for those not wanting to use auto-install?
>> --
>> You received this message because you are subscribed to the Google Groups
>> "Jenkins Users" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected].
>> For more options, visit https://groups.google.com/d/optout.
>>
>>
>>
>>
>> -------------------------------------------------------------------------------
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "Jenkins Users" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected].
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>
>
> --
> Jeff Vincent
> See my LinkedIn profile at:
> http://www.linkedin.com/in/rjeffreyvincent
>
> --
> You received this message because you are subscribed to the Google Groups
> "Jenkins Users" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/d/optout.
>



-- 
Thanks!
Mark Waite

-- 
You received this message because you are subscribed to the Google Groups 
"Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to