Selecting the best performing node in a label

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Selecting the best performing node in a label

terrapin
Suppose a project is assigned to a label and it turns out that this
project performs better (=faster) on some nodes (on a historical
basis) than others. Is there any possibility, means that Jenkins will
then select the better performing nodes when assigning the job to a
node?
Reply | Threaded
Open this post in threaded view
|

Re: Selecting the best performing node in a label

Mark Waite
I've not heard anything more about the scheduling algorithm than is described on the wiki and summarized at http://jenkins.361315.n4.nabble.com/Hudson-scheduling-strategy-tc3046471.html.  There isn't any selection condition known to me which controls slave selection based on node performance metrics of any sort.

Mark Waite


From: terrapin <[hidden email]>
To: Jenkins Users <[hidden email]>
Sent: Tue, March 8, 2011 3:38:52 AM
Subject: Selecting the best performing node in a label

Suppose a project is assigned to a label and it turns out that this
project performs better (=faster) on some nodes (on a historical
basis) than others. Is there any possibility, means that Jenkins will
then select the better performing nodes when assigning the job to a
node?
Reply | Threaded
Open this post in threaded view
|

Re: Selecting the best performing node in a label

Les Mikesell
In reply to this post by terrapin
On 3/8/2011 4:38 AM, terrapin wrote:
> Suppose a project is assigned to a label and it turns out that this
> project performs better (=faster) on some nodes (on a historical
> basis) than others. Is there any possibility, means that Jenkins will
> then select the better performing nodes when assigning the job to a
> node?

If you know some nodes have better performance you can set a higher
number of executors for them which should increase their odds of getting
a job.

--
   Les Mikesell
    [hidden email]