[ClusterLabs] Antw: Re: Antw: Re: Antw: Pacemaker 1.1.16 - Release Candidate 1

[Vladislav Bogdanov]

09.11.2016 10:59, Ulrich Windl wrote:
>>>> Ken Gaillot <kgaillot at redhat.com> schrieb am 08.11.2016 um 18:16 in Nachricht
> <92c4a0de-33ce-cdc2-a778-17fddfe63865 at redhat.com>:
>> On 11/08/2016 03:02 AM, Ulrich Windl wrote:
>
> [...]
>> The user is responsible for choosing meaningful values. For example, if
>> node-health-base is +10 but yellow is -15, then any yellow attribute
>> will still push resources away. Of course, that could still be
>> meaningful when combined with other scores -- someone might do that if
>> they want a location preference of +5 to counteract a single yellow
>> attribute. Or maybe instead of node-health-base, someone sets a positive
>> stickiness, so existing resources can stay on a yellow node, but new
>> resources won't be placed there. It can be as simple or complicated as
>> you want to get :)
>
> I think it's too complicated, already: In my simple world nodes with
> status "green" are OK to run any resource, nodes with status "yellow"
> should not start new resources, and nodes with status "red" should move
> away running resources. Ok, I see the some people have a desire for
> "orange", "indian yellow" and a lots of different colors on the
> spectrum, but thinking of the seemingly endless cases to test, I prefer
> a simple world ;-)

That was implemented mainly to allow drbd (among other replicated 
datastores) to be started as a secondary on a node with yellow 
attributes. That attributes come from outside of pacemaker and actually 
may have any meaning one may imagine. In our case we utilized new alerts 
feature to manage specific yellow attributes if resources repeatedly 
fail and recover, something similar to a flapping detection in nagios. 
So all services go away from a node with the worst health score, making 
another node "active" one. But we still wanted to have secondary data 
replica online and in sync.
The first attempt was to set positive "base" health attribute from a 
resource agent, but then we realized that resource won't start (rather 
expected, like every other resource) after node with some yellow 
attributes was brought back from the standby mode.
So the decision was made to submit a 5-line patch for pacemaker ;)

>
>>
>> _______________________________________________
>> Users mailing list: Users at clusterlabs.org
>> http://clusterlabs.org/mailman/listinfo/users
>>
>> Project Home: http://www.clusterlabs.org
>> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
>> Bugs: http://bugs.clusterlabs.org
>
>
>
>
>
> _______________________________________________
> Users mailing list: Users at clusterlabs.org
> http://clusterlabs.org/mailman/listinfo/users
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
>