[Pacemaker] RFC: What part of the XML configuration do you hate themost?

[Andrew Beekhof]

On May 8, 2008, at 7:52 AM, Junko IKEDA wrote:

> Hi,
>
>> The changes for Pacemaker 1.0 include an overhaul of configuration
>> syntax.
>>
>> We have a few things in mind for this already, but I'd also like to
>> get people's opinion on which parts need the most attention.
>
> When are you planning to release Pacemaker 1.0?

1.0 will probably be out some time around August or September.

HOWEVER, the features will be available in the 0.7 series of  
developmental* releases (starting later this month) so we can get  
feedback on bugs and the general usability of new features.

* Developmental in the sense that:
    - there will be new features added (or possibly removed)
    - the behavior of features may change
    - the configuration syntax may change
    - various features will generally not be stable or complete
    - there will probably be more bugs than usual

Which can sound a bit scary, but the more people try it out, the more  
likely things will be usable, functional and bug-free when we lock  
everything down for the 1.0 stable series.

>
> New features which you are implementing in dev tree are very  
> attractive.
> Especially, "migration-threshold" is what our customer desires.
> They get bored with handling some complex score calculation...

Yeah, it seemed like a good idea at the time but it has been clear for  
a while that it was a huge mistake.

> If we set migration-threshold="3", the resource will fail over after  
> three
> failures, right?

Exactly.

> and I'm trying "expired fail-count" now,
> It seems that something wrong with getting time value for fail count.

ooops.
applied, thanks!