Selecting nodes

Every command that acts on nodes takes a node set. The syntax is ClusterShell’s, so it is what you already type at nodeset, clush and sinfo.

The syntax

$ clusterctl node select 'exe0001'                 # one host
$ clusterctl node select 'exe[1-10]'               # a range
$ clusterctl node select 'exe[0001-0010]'          # padded
$ clusterctl node select 'exe[1-10/2]'             # with a step
$ clusterctl node select 'exe[1,5,9-12]'           # several ranges
$ clusterctl node select 'rack[1-2]node[01-04]'    # two dimensions
$ clusterctl node select '@compute'                # a group
$ clusterctl node select '@slurm:idle'             # a group from a source

Combining sets

Operator	Meaning	Example
`,` or a space	union	`exe[1-4],sub[1-2]`
`!`	difference	`@compute!@drained`
`&`	intersection	`@idle&@rack:R02`
`^`	in exactly one	`@yesterday^@today`

Operators have no precedence. An expression is evaluated strictly left to right, so @a!@b&@c is ((@a minus @b) intersect @c). Write the order you mean; there are no brackets for grouping.

Folding and expanding

$ clusterctl node select 'exe[1-3],exe7'
exe[0001-0003,0007]

$ clusterctl node select 'exe[1-3]' --expand
exe0001
exe0002
exe0003

$ clusterctl node select '@compute' --count
1024

Note the padding. The inventory wrote these machines as exe0001, so a set comes back under the names the site gave them even when you typed exe1. exe1 and exe0001 are the same host: padding is how a name is displayed, not part of what it identifies.

Groups

A group is resolved by a source. Which sources exist is configuration:

$ clusterctl node groups
SOURCE     GROUP
inventory  exe
inventory  sub
inventory  wlm
rack       R01
rack       R02
slurm      main
slurm      debug

A bare @group searches the default source first and then the others, so @exe works without a prefix. @source:group names one explicitly, which is what to write when two sources could both answer.

$ clusterctl node groups exe0007
SOURCE     GROUPS
inventory  exe
rack       R02
slurm      main

Where groups come from

groups:
  defaultSource: inventory
  sources:
    inventory:
      # One group per value of a node attribute. This is what a genders
      # file provided: @exe, @wlm, @dbm.
      attribute: class

    rack:
      attribute: rack

    static:
      # A table. A group may refer to other groups.
      static:
        infra: wlm01,dbm01
        compute: "@inventory:exe"

    slurm:
      # Asked of the workload manager and cached for a minute.
      cacheTtl: 60s
      exec:
        role: login
        map: [sinfo, -h, -o, "%N", -p, $GROUP]
        all: [sinfo, -h, -o, "%N"]
        list: [sinfo, -h, -o, "%R"]
        reverse: [sinfo, -h, -N, -o, "%R", -n, $NODE]

An exec source is given an argument vector, and $GROUP and $NODE are substituted as whole arguments. A group name is never interpreted by a shell.

A selection for a whole session

$ export CLUSTERCTL_NODES='@rack:R02'
$ clusterctl bmc status
$ clusterctl exec -- uptime

-n always wins over it. Nothing else falls back to a set you set earlier: a command with no selection stops rather than guessing.

Passing a set to another tool

$ clusterctl slurm node nodeset drain
exe[0007,0042,0511]

$ clusterctl exec -n "$(clusterctl slurm node nodeset idle)" -- uptime

Slurm and FreeIPMI understand one bracketed range per name and neither several numeric dimensions nor groups, so clusterctl expands a set for them when it has to. You never have to think about it.

Checking a set before you use it

$ clusterctl node select '@rack:R02&@slurm:idle' --expand
$ clusterctl node list '@rack:R02'
$ clusterctl node fqdn -n '@rack:R02'

Worth doing before anything destructive. --dry-run also prints the set it resolved before it stops.

Running commands