First commands
Everything here is read-only. Run it in order the first time you point clusterctl at a site.
Does the configuration resolve
$ clusterctl config validate
FILE KIND NAME
/etc/clusterctl/cluster.yaml Cluster cluster1
/etc/clusterctl/config.yaml Config
/etc/clusterctl/inventory.yaml NodeInventory example
/etc/clusterctl/site.yaml Site example
4 documents are valid; context cluster1 resolves
A mistake is reported where you wrote it:
$ clusterctl config validate
clusterctl: site.yaml is not valid:
site.yaml:9:7: spec.hosts.login.forwardAgnet: unknown field "forwardAgnet"; did you mean "forwardAgent"?
Can it reach things
$ clusterctl doctor
CHECK STATUS DETAIL
configuration ok context cluster1, cluster cluster1, site example
ssh client ok /usr/bin/ssh OpenSSH_9.6p1
generated ssh configuration ok ~/.local/state/clusterctl/ssh_config
host key file ok 412 entries in /etc/clusterctl/ssh-known-hosts
node inventory ok 1024 nodes
$ clusterctl doctor --remote
--remote contacts each configured host role and checks the programs the
commands need are there.
Do node sets resolve
$ clusterctl node select '@compute'
exe[0001-1024]
$ clusterctl node select '@compute!@drained' --count
1019
$ clusterctl node groups
SOURCE GROUP
inventory exe
inventory sub
rack R01
slurm main
What is known about a node
$ clusterctl node describe exe0007
FIELD VALUE
name exe0007
host exe0007.hpc.example.org
bmc exe0007.mgmt.hpc.example.org
address 10.0.2.7
rack R02
attribute.class exe
attribute.vendor vendor2
groups.inventory exe
groups.rack R02
Can it reach a host
$ clusterctl login --dry-run login
ssh -F ~/.local/state/clusterctl/ssh_config alice_adm@login.hpc.example.org
$ clusterctl login login -- uptime
14:42:30 up 31 days, 4:10, 0 users, load average: 12.04, 11.87, 11.90
Can it reach the nodes
$ clusterctl exec -n 'exe[1-4]' -- uptime -p
exe0001: up 3 weeks, 1 day
exe0002: up 3 weeks, 1 day
exe0003: up 3 weeks, 1 day
exe0004: up 2 days
$ clusterctl exec -n '@compute' --dedup -- uname -r
exe[0001-1020] (1020)
5.14.0-570.el9.x86_64
exe[1021-1024] (4)
5.14.0-503.el9.x86_64
What does the hardware say
$ clusterctl bmc status -n 'exe[1-4]'
NODE STATE ERROR
exe0001 On
exe0002 On
exe0003 Off
exe0004 On
If this is the first time clusterctl has spoken to those service processors,
it records the certificate each one presents. A later change is refused until
you say the replacement was intentional with clusterctl bmc forget.