service data: only set failed_nodes key if needed

[pve-ha-manager.git] / src / PVE / HA / Manager.pm

diff --git a/src/PVE/HA/Manager.pm b/src/PVE/HA/Manager.pm
index 2933926ae47ed49006d743bfd7702d3a3577bc13..0b90de4baf1c5e0b417472c618cb3aae8af4d868 100644 (file)
--- a/src/PVE/HA/Manager.pm
+++ b/src/PVE/HA/Manager.pm
@@ -43,9 +43,9 @@ sub flush_master_status {
     $ms->{node_status} = $ns->{status};
     $ms->{service_status} = $ss;
     $ms->{timestamp} = $haenv->get_time();
-    
+
     $haenv->write_manager_status($ms);
-} 
+}
 
 sub get_service_group {
     my ($groups, $online_node_usage, $service_conf) = @_;
@@ -57,7 +57,7 @@ sub get_service_group {
     }
 
     # overwrite default if service is bound to a specific group
-    $group =  $groups->{ids}->{$service_conf->{group}} if $service_conf->{group} && 
+    $group =  $groups->{ids}->{$service_conf->{group}} if $service_conf->{group} &&
        $groups->{ids}->{$service_conf->{group}};
 
     return $group;
@@ -118,7 +118,7 @@ sub select_service_node {
        }
     }
 
-    my @nodes = sort { 
+    my @nodes = sort {
        $online_node_usage->{$a} <=> $online_node_usage->{$b} || $a cmp $b
     } keys %{$pri_groups->{$top_pri}};
 
@@ -152,7 +152,7 @@ my $uid_counter = 0;
 
 sub compute_new_uuid {
     my ($state) = @_;
-    
+
     $uid_counter++;
     return md5_base64($state . $$ . time() . $uid_counter);
 }
@@ -183,7 +183,7 @@ sub recompute_online_node_usage {
        my $sd = $self->{ss}->{$sid};
        my $state = $sd->{state};
        if (defined($online_node_usage->{$sd->{node}})) {
-           if (($state eq 'started') || ($state eq 'request_stop') || 
+           if (($state eq 'started') || ($state eq 'request_stop') ||
                ($state eq 'fence') || ($state eq 'freeze') || ($state eq 'error')) {
                $online_node_usage->{$sd->{node}}++;
            } elsif (($state eq 'migrate') || ($state eq 'relocate')) {
@@ -218,7 +218,7 @@ my $change_service_state = sub {
 
     $sd->{state} = $new_state;
     $sd->{node} = $old_node;
-    $sd->{failed_nodes} = $old_failed_nodes;
+    $sd->{failed_nodes} = $old_failed_nodes if defined($old_failed_nodes);
 
     my $text_state = '';
     foreach my $k (sort keys %params) {
@@ -243,7 +243,7 @@ my $fence_recovery_cleanup = sub {
 
     my $haenv = $self->{haenv};
 
-    my (undef, $type, $id) = PVE::HA::Tools::parse_sid($sid);
+    my (undef, $type, $id) = $haenv->parse_sid($sid);
     my $plugin = PVE::HA::Resources->lookup($type);
 
     # should not happen
@@ -299,7 +299,7 @@ my $recover_fenced_service = sub {
     }
 };
 
-# read LRM status for all nodes 
+# read LRM status for all nodes
 sub read_lrm_status {
     my ($self) = @_;
 
@@ -317,7 +317,7 @@ sub read_lrm_status {
        }
     }
 
-    
+
     return ($results, $modes);
 }
 
@@ -328,20 +328,19 @@ sub update_crm_commands {
     my ($haenv, $ms, $ns, $ss) = ($self->{haenv}, $self->{ms}, $self->{ns}, $self->{ss});
 
     my $cmdlist = $haenv->read_crm_commands();
-    return if !defined($cmdlist);
 
     foreach my $cmd (split(/\n/, $cmdlist)) {
        chomp $cmd;
 
        if ($cmd =~ m/^(migrate|relocate)\s+(\S+)\s+(\S+)$/) {
-           my ($task, $sid, $node) = ($1, $2, $3); 
+           my ($task, $sid, $node) = ($1, $2, $3);
            if (my $sd = $ss->{$sid}) {
                if (!$ns->node_is_online($node)) {
                    $haenv->log('err', "crm command error - node not online: $cmd");
                } else {
                    if ($node eq $sd->{node}) {
                        $haenv->log('info', "ignore crm command - service already on target node: $cmd");
-                   } else { 
+                   } else {
                        $haenv->log('info', "got crm command: $cmd");
                        $ss->{$sid}->{cmd} = [ $task, $node];
                    }
@@ -405,7 +404,7 @@ sub manage {
 
     for (;;) {
        my $repeat = 0;
-       
+
        $self->recompute_online_node_usage();
 
        foreach my $sid (sort keys %$ss) {
@@ -559,7 +558,7 @@ sub next_state_stopped {
        # this can happen if we fence a node with active migrations
        # hack: modify $sd (normally this should be considered read-only)
        $haenv->log('info', "fixup service '$sid' location ($sd->{node} => $cd->{node})");
-       $sd->{node} = $cd->{node}; 
+       $sd->{node} = $cd->{node};
     }
 
     if ($sd->{cmd}) {
@@ -577,7 +576,7 @@ sub next_state_stopped {
                return;
            }
        } else {
-           $haenv->log('err', "unknown command '$cmd' for service '$sid'"); 
+           $haenv->log('err', "unknown command '$cmd' for service '$sid'");
        }
     }
 
@@ -631,7 +630,7 @@ sub next_state_started {
        }
        return;
     }
-       
+
     if ($cd->{state} eq 'disabled' || $cd->{state} eq 'stopped') {
        &$change_service_state($self, $sid, 'request_stop');
        return;
@@ -653,7 +652,7 @@ sub next_state_started {
                    &$change_service_state($self, $sid, $cmd, node => $sd->{node}, target => $target);
                }
            } else {
-               $haenv->log('err', "unknown command '$cmd' for service '$sid'"); 
+               $haenv->log('err', "unknown command '$cmd' for service '$sid'");
            }
        } else {