* postservice/31-one-netmng (OneClient.create_network): make sure to
return True since the returned value is checkd in
“OneNetwork.create()”.
Ref: #24531
* dicos/29_one-master.xml: “OpenNebulaNetwork.conf” is moved for 5.6.
* tmpl/OpenNebulaNetwork.conf: update to match 5.5.80 configuration.
* tmpl/ldap_auth.conf: ditoo
* tmpl/oned.conf: ditoo.
* tmpl/vmm_exec_kvm.conf: ditoo.
Ref: #24197
* dicos/29_one-master.xml: “one_cluster_name” store the name of the
default cluster.
* postservice/30-one-mng: take care of the customizable cluster name.
* postservice/31-one-netmng: ditoo.
* scripts/onehost_create_all: ditoo.
Ref: #23553
The “libvirt-guests” service can conflict with “onevm-all” and we
don't need to wait for each VM to boot during start.
* posttemplate/10-libvirt-guests: disable the “libvirt-guests” service.
* init/onenode.service (ExecStart): use default wait timeout (60s).
(TimeoutStartSec): wait for longer than the default timeout.
(ExecReload): just try to resume any remaining VMs.
(ExecStop): wait longer for VM to suspend.
(TimeoutStopSec): wait for longer than the stop timeout.
* scripts/onevm-all: schedule actions in parallel and wait globally
for their executions.
Ref: #22155
If a user suspend a VM manually, onenode.service must not resume it
automatically.
* scripts/onevm-all: Check if suspended VMs are in the saved list of
running VMs.
Ref: #20718
Le service onenode ne se lance pas a cause d'une
boucle dans les dépendances de services.
Pour régler le problème on le lance après la multi-user.target
De plus au moment ou le service se lance opennebula n'est pas
complètement lancé alors le script essaye d'ouvrir des connections
pendant 20 secondes avant de remonter un problème.
Enfin on garde une liste des machines qui sont "running" avant l'arrêt
du serveur pour pouvoir les relance proprement.
Contribution de Cadoles (htts://www.cadoles.com)
ref #20338 @6h
vincent
Emmanuel Garette
Philippe Caseiro
Daniel Dehennin