[ClusterLabs] Problems with master/slave failovers

[Andrei Borzenkov]

29.06.2019 8:05, Harvey Shepherd пишет:
> There is an ordering constraint - everything must be started after the king resource. But even if this constraint didn't exist I don't see that it should logically make any difference due to all the non-clone resources being colocated with the master of the king resource. Surely it would make no sense for Pacemaker to start or move colocated resources until a master king resource has been elected?
> 
>   <tags>
>     <tag id="servant2_dependents">
>       <obj_ref id="servant4"/>
>       <obj_ref id="servant5"/>
>       <obj_ref id="servant6"/>
>       <obj_ref id="servant7"/>
>       <obj_ref id="servant8"/>
>       <obj_ref id="servant9_active_disabled"/>
>       <obj_ref id="servant11"/>
>       <obj_ref id="servant12"/>
>       <obj_ref id="servant13"/>
>     </tag>
>   </tags>
>   <constraints>
>     <rsc_colocation id="colocation_with_king_resource_master" score="INFINITY">
>       <resource_set id="king_resource_master_dependents" sequential="false">
>         <resource_ref id="stk_shared_ip"/>
>         <resource_ref id="servant4"/>
>         <resource_ref id="servant5"/>
>         <resource_ref id="servant6"/>
>         <resource_ref id="servant7"/>
>         <resource_ref id="servant8"/>
>         <resource_ref id="servant9_active_disabled"/>
>         <resource_ref id="servant10"/>
>         <resource_ref id="servant11"/>
>         <resource_ref id="servant12"/>
>         <resource_ref id="servant13"/>
>       </resource_set>
>       <resource_set id="king_resource_master" sequential="true" role="Master">
>         <resource_ref id="ms_king_resource"/>
>         <resource_ref id="ms_servant2"/>
>         <resource_ref id="ms_servant3"/>
>       </resource_set>
>     </rsc_colocation>
>     <rsc_order id="dependents_after_servant2" kind="Mandatory" first="ms_servant2" then="servant2_dependents"/>
>   </constraints>
> 

This ordering constraint is satisfied by slave of ms_servant2. Slave is
already started at the point failover happens so pacemaker is free to
start all other resources immediately. If you intend to order against
master, you need first-action="promote" then-action="start".