[Pacemaker] [Question] About replacing in resource_set of the order limitation.

[renayama19661014 at ybb.ne.jp]

Hi Andrew,

Thank you for comments.

> Is this related to your email about symmetrical not being defaulted consistently between colocate_rsc_sets() and unpack_colocation_set()?

Yes.
I think that a default is not handled well.
I will not have any problem when "sequential" attribute is set in cib by all means.

I think that I should revise processing when "sequential" attribute is not set.

Best Regards,
Hideo Yamauchi.

> 
> On 22 Jan 2014, at 3:05 pm, renayama19661014 at ybb.ne.jp wrote:
> 
> > Hi All,
> > 
> > My test seemed to include a mistake.
> > It seems to be replaced by two limitation.
> > 
> >> However, I think that symmetircal="false" is applied to all order limitation in this.
> >> (snip)
> >>       <rsc_order id="rsc_order-clnPing-grpPg1" score="0" symmetrical="false">
> >>         <resource_set id="rsc_order-clnPing-grpPg1-0">
> >>           <resource_ref id="clnPing"/>
> >>         </resource_set>
> >>         <resource_set id="rsc_order-clnPing-grpPg1-1".....>
> >>           <resource_ref id="A"/>
> >>           .......
> >>           <resource_ref id="F"/>
> >>         </resource_set>
> >>       </rsc_order>
> >> (snip)
> > 
> > 
> >      <rsc_order id="rsc_order-clnPing-grpPg1" score="0" first="clnPing" then="prmEx" symmetrical="false">
> >      </rsc_order>
> >      <rsc_order id="rsc_order-clnPing-grpPg2" score="0" symmetrical="true">
> >        <resource_set id="rsc_order-clnPing-grpPg2-0" require-all="false">
> >          <resource_ref id="prmEx"/>
> >          <resource_ref id="prmFs1"/>
> >          <resource_ref id="prmFs2"/>
> >          <resource_ref id="prmFs3"/>
> >          <resource_ref id="prmIp"/>
> >          <resource_ref id="prmPg"/>
> >        </resource_set>
> >      </rsc_order>
> > 
> > If my understanding includes a mistake, please point it out.
> > 
> > Best Reagards,
> > Hideo Yamauchi.
> > 
> > --- On Fri, 2014/1/17, renayama19661014 at ybb.ne.jp <renayama19661014 at ybb.ne.jp> wrote:
> > 
> >> Hi All,
> >> 
> >> We confirm a function of resource_set.
> >> 
> >> There were the resource of the group and the resource of the clone.
> >> 
> >> (snip)
> >> Stack: corosync
> >> Current DC: srv01 (3232238180) - partition WITHOUT quorum
> >> Version: 1.1.10-f2d0cbc
> >> 1 Nodes configured
> >> 7 Resources configured
> >> 
> >> 
> >> Online: [ srv01 ]
> >> 
> >> Resource Group: grpPg
> >>      A      (ocf::heartbeat:Dummy): Started srv01 
> >>      B      (ocf::heartbeat:Dummy): Started srv01 
> >>      C      (ocf::heartbeat:Dummy): Started srv01 
> >>      D      (ocf::heartbeat:Dummy): Started srv01 
> >>      E      (ocf::heartbeat:Dummy): Started srv01 
> >>      F      (ocf::heartbeat:Dummy): Started srv01 
> >> Clone Set: clnPing [prmPing]
> >>      Started: [ srv01 ]
> >> 
> >> Node Attributes:
> >> * Node srv01:
> >>     + default_ping_set                  : 100       
> >> 
> >> Migration summary:
> >> * Node srv01: 
> >> 
> >> (snip)
> >> 
> >> These have limitation showing next.
> >> 
> >> (snip)
> >>       <rsc_colocation id="rsc_colocation-grpPg-clnPing" score="INFINITY" rsc="grpPg" with-rsc="clnPing">
> >>       </rsc_colocation>
> >>       <rsc_order id="rsc_order-clnPing-grpPg" score="0" first="clnPing" then="grpPg" symmetrical="false">
> >>       </rsc_order>
> >> (snip)
> >> 
> >> 
> >> We tried that we rearranged a group in resource_set.
> >> I think that I can rearrange the limitation of "colocation" as follows.
> >> 
> >> (snip)
> >>       <rsc_colocation id="rsc_colocation-grpPg-clnPing" score="INFINITY">
> >>         <resource_set id="rsc_colocation-grpPg-clnPing-0">
> >>           <resource_ref id="clnPing"/>
> >>           <resource_ref id="A"/>
> >>           .......
> >>           <resource_ref id="F"/>
> >>         </resource_set>
> >>       </rsc_colocation>
> >> (snip)
> >> 
> >> How should I rearrange the limitation of "order" in resource_set?
> >> 
> >> I thought that it was necessary to list two of the next, but a method to express well was not found.
> >> 
> >> * "symmetirical=true" is necessary between the resources that were a group(A to F).
> >> * "symmetirical=false" is necessary between the resource that was a group(A to F) and the clone resources.
> >> 
> >> I wrote it as follows.
> >> However, I think that symmetircal="false" is applied to all order limitation in this.
> >> (snip)
> >>       <rsc_order id="rsc_order-clnPing-grpPg1" score="0" symmetrical="false">
> >>         <resource_set id="rsc_order-clnPing-grpPg1-0">
> >>           <resource_ref id="clnPing"/>
> >>         </resource_set>
> >>         <resource_set id="rsc_order-clnPing-grpPg1-1".....>
> >>           <resource_ref id="A"/>
> >>           .......
> >>           <resource_ref id="F"/>
> >>         </resource_set>
> >>       </rsc_order>
> >> (snip)
> >> 
> >> Best Reards,
> >> Hideo Yamauchi.
> >> 
> >> 
> >> 
> >> _______________________________________________
> >> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> >> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> >> 
> >> Project Home: http://www.clusterlabs.org
> >> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> >> Bugs: http://bugs.clusterlabs.org
> >> 
> > 
> > _______________________________________________
> > Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> > http://oss.clusterlabs.org/mailman/listinfo/pacemaker
> > 
> > Project Home: http://www.clusterlabs.org
> > Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> > Bugs: http://bugs.clusterlabs.org
> 
>