The minimum value for timeout in vm_shutdown is changed from 0 to 1, since a
value of 0 would trigger a hard stop for HA managed VMs. Like this the API
description stays valid for all cases.
Signed-off-by: Fabian Ebner <f.eb...@proxmox.com>
---
In vm_shutdown we'd like to pass along the timeout parameter to the HA stack,
but the value 0 triggers a special behavior there, namely a hard stop.
So either the description needs to be changed to mention this behavior or
we just don't allow a timeout of 0 in vm_shutdown. Since a shutdown with
timeout 0 doesn't really make sense anyways, I opted for the latter.
It's the same situation for containers.
PVE/API2/Qemu.pm | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/PVE/API2/Qemu.pm b/PVE/API2/Qemu.pm
index c31dd1d..16a7a04 100644
--- a/PVE/API2/Qemu.pm
+++ b/PVE/API2/Qemu.pm
@@ -2111,7 +2111,7 @@ __PACKAGE__->register_method({
print "Requesting HA stop for VM $vmid\n";
- my $cmd = ['ha-manager', 'set', "vm:$vmid", '--state',
'stopped'];
+ my $cmd = ['ha-manager', 'crm-command', 'stop', "vm:$vmid",
'0'];
PVE::Tools::run_command($cmd);
return;
};
@@ -2204,7 +2204,7 @@ __PACKAGE__->register_method({
timeout => {
description => "Wait maximal timeout seconds.",
type => 'integer',
- minimum => 0,
+ minimum => 1,
optional => 1,
},
forceStop => {
@@ -2267,12 +2267,13 @@ __PACKAGE__->register_method({
if (PVE::HA::Config::vm_is_ha_managed($vmid) && $rpcenv->{type} ne
'ha') {
+ my $timeout = $param->{timeout} // 60;
my $hacmd = sub {
my $upid = shift;
print "Requesting HA stop for VM $vmid\n";
- my $cmd = ['ha-manager', 'set', "vm:$vmid", '--state',
'stopped'];
+ my $cmd = ['ha-manager', 'crm-command', 'stop', "vm:$vmid",
"$timeout"];
PVE::Tools::run_command($cmd);
return;
};
--
2.20.1
_______________________________________________
pve-devel mailing list
pve-devel@pve.proxmox.com
https://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
