proxmox-mirrors/pve-manager
1
0
Fork 0
mirror of https://git.proxmox.com/git/pve-manager synced 2025-08-15 05:25:49 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

pvescheduler: fix potential stall on full shutdown

Browse Source 
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
This commit is contained in:
Thomas Lamprecht 2021-11-23 08:58:16 +01:00
parent 8f3b5bbb91
commit 9c1943935c
1 changed files with 14 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
PVE/Service/pvescheduler.pm
View File

@ -19,6 +19,12 @@ my $daemon = __PACKAGE__->new('pvescheduler', $cmdline, %daemon_options);
my @JOB_TYPES = qw(replication jobs);
my sub running_job_pids : prototype($) {
my ($self) = @_;
my $pids = [ map { keys $_->%* } values $self->{jobs}->%* ];
return scalar($pids->@*) ? $pids : undef;
}
my $finish_jobs = sub {
my ($self) = @_;
for my $type (@JOB_TYPES) {
@ -134,21 +140,19 @@ sub run {
# NOTE: we only get here on shutdown_request, so we already sent a TERM to all job-types
my $timeout = 0;
while (scalar(keys $jobs->%*)) {
for my $type (keys $jobs->%*) {
next if !scalar(keys $jobs->{$type}->%*);
kill 'TERM', keys $jobs->{$type}->%*;
}
$finish_jobs->($self); # doesn't hurt
while(my $pids = running_job_pids($self)) {
kill 'TERM', $pids->@*; # send TERM to all workers at once, possible thundering herd - FIXME?
$finish_jobs->($self);
# some jobs have a lock timeout of 60s, wait a bit more for graceful termination
last if $timeout > 75;
$timeout += sleep(3);
}
for my $type (keys $jobs->%*) { # ensure the rest gets stopped
my @pids = keys $jobs->{$type}->%*;
syslog('warn', "unresponsive job-worker, killing now: " . join(', ', @pids));
kill 'KILL', @pids;
if (my $pids = running_job_pids($self)) {
syslog('warn', "unresponsive job-worker, killing now: " . join(', ', $pids->@*));
kill 'KILL', $pids->@*;
}
}