[ClusterLabs] crm shell 2.1.2 manual bug?

[Andrei Borzenkov]

28.03.2018 13:25, Ulrich Windl пишет:
> Hi!
> 
> For crmsh-2.1.2+git132.gbc9fde0-18.2 I think there's a bug in the manual describing resource sets:
> 
>        sequential
>            If true, the resources in the set do not depend on each other internally. Setting sequential to true implies a strict order of dependency within the set.
> 
> Obviously "true" cannot mean both: "do not depend" and "depend". My guess is that the first true has to be false.
> 
> In addition, isn't the syntax with resource sets missing something (despite from the description)?:
> 
> colocation <id> <score>: resource_sets
>                  [node-attribute=<node_attr>]
> 
> comparing to
> 
> colocation <id> <score>: <rsc>[:<role>] <with-rsc>[:<role>]
>                  [node-attribute=<node_attr>]
> 
> I expected the with-rsc to be at the end also when using resource sets...
> 

No, colocating sets works indeed differently, syntax is correct (for XML
examples in pacemaker explained that is ...)

> I came across this when trying to add a colocation like this:
> colocation col_LV inf:( cln_LV cln_LV-L1 cln_LV-L2 cln_ML cln_ML-L1 cln_ML-L2 ) cln_VMs
> 
> crm complained about this:
> ERROR: 1: syntax in role: Unmatched opening bracket near <cln_ML-L1> parsing 'colocation ...'
> ERROR: 2: syntax: Unknown command near <cln_ml-l2> parsing 'cln_ml-l2 ) cln_VMs'
> (note the lower case)
> 

You need to create set consisting of single resource. I do not know if
it possible to colocate set and single resource - none of examples show it.

See also
https://clusterlabs.org/pacemaker/doc/en-US/Pacemaker/1.1/html/Pacemaker_Explained/s-resource-sets-colocation.html