use Data::Dumper;
use JSON;
use Time::HiRes qw(gettimeofday tv_interval);
+use POSIX qw(strftime);
use PVE::INotify;
use PVE::ProcFSTools;
# regression tests should overwrite this
sub get_log_time {
- return time();
+ return strftime("%F %H:%M:%S", localtime);
}
# Find common base replication snapshot, available on local and remote side.
$last_snapshots->{$volid}->{$snap} = 1;
} elsif ($snap =~ m/^\Q$prefix\E/) {
$logfunc->("delete stale replication snapshot '$snap' on $volid");
- PVE::Storage::volume_snapshot_delete($storecfg, $volid, $snap);
- $cleaned_replicated_volumes->{$volid} = 1;
+
+ eval {
+ PVE::Storage::volume_snapshot_delete($storecfg, $volid, $snap);
+ $cleaned_replicated_volumes->{$volid} = 1;
+ };
+
+ # If deleting the snapshot fails, we can not be sure if it was due to an error or a timeout.
+ # The likelihood that the delete has worked out is high at a timeout.
+ # If it really fails, it will try to remove on the next run.
+
+ # warn is for syslog/journal.
+ warn $@ if $@;
+
+ # logfunc will written in replication log.
+ $logfunc->("delete stale replication snapshot error: $@") if $@;
}
}
}
}
sub replicate_volume {
- my ($ssh_info, $storecfg, $volid, $base_snapshot, $sync_snapname, $rate, $insecure) = @_;
+ my ($ssh_info, $storecfg, $volid, $base_snapshot, $sync_snapname, $rate, $insecure, $logfunc) = @_;
my ($storeid, $volname) = PVE::Storage::parse_volume_id($volid);
my $ratelimit_bps = int(1000000*$rate) if $rate;
PVE::Storage::storage_migrate($storecfg, $volid, $ssh_info, $storeid, $volname,
- $base_snapshot, $sync_snapname, $ratelimit_bps, $insecure);
+ $base_snapshot, $sync_snapname, $ratelimit_bps, $insecure, 1, $logfunc);
}
if $start_time <= $last_sync;
my $vmid = $jobcfg->{guest};
- my $vmtype = $jobcfg->{vmtype};
my $conf = $guest_class->load_config($vmid);
my ($running, $freezefs) = $guest_class->__snapshot_check_freeze_needed($vmid, $conf, 0);
$running //= 0; # to avoid undef warnings from logfunc
- $logfunc->("guest => $vmid, type => $vmtype, running => $running");
+ my $guest_name = $guest_class->guest_type() . ' ' . $vmid;
+
+ $logfunc->("guest => $guest_name, running => $running");
$logfunc->("volumes => " . join(',', @$sorted_volids));
if (my $remove_job = $jobcfg->{remove_job}) {
};
my $err = $@;
- # unfreeze immediately
+ # thaw immediately
if ($freezefs) {
+ $logfunc->("thaw guest filesystem");
$guest_class->__snapshot_freeze($vmid, 1);
}
$logfunc->("full sync '$volid' ($sync_snapname)");
}
- replicate_volume($ssh_info, $storecfg, $volid, $base_snapname, $sync_snapname, $rate, $insecure);
+ replicate_volume($ssh_info, $storecfg, $volid, $base_snapname, $sync_snapname, $rate, $insecure, $logfunc);
}
};
$err = $@;
# remove old snapshots because they are no longer needed
$cleanup_local_snapshots->($last_snapshots, $last_sync_snapname);
- remote_finalize_local_job($ssh_info, $jobid, $vmid, $sorted_volids, $start_time, $logfunc);
+ eval {
+ remote_finalize_local_job($ssh_info, $jobid, $vmid, $sorted_volids, $start_time, $logfunc);
+ };
- die $err if $err;
+ # old snapshots will removed by next run from prepare_local_job.
+ if ($err = $@) {
+ # warn is for syslog/journal.
+ warn $err;
+
+ # logfunc will written in replication log.
+ $logfunc->("delete stale replication snapshot error: err");
+ }
return $volumes;
}
my $run_replication_nolock = sub {
- my ($guest_class, $jobcfg, $iteration, $start_time, $logfunc) = @_;
+ my ($guest_class, $jobcfg, $iteration, $start_time, $logfunc, $verbose) = @_;
my $jobid = $jobcfg->{id};
# we normaly write errors into the state file,
# but we also catch unexpected errors and log them to syslog
# (for examply when there are problems writing the state file)
- eval {
- my $state = PVE::ReplicationState::read_job_state($jobcfg);
-
- PVE::ReplicationState::record_job_start($jobcfg, $state, $start_time, $iteration);
- my $t0 = [gettimeofday];
+ my $state = PVE::ReplicationState::read_job_state($jobcfg);
- mkdir $PVE::ReplicationState::replicate_logdir;
- my $logfile = PVE::ReplicationState::job_logfile_name($jobid);
- open(my $logfd, '>', $logfile) ||
- die "unable to open replication log '$logfile' - $!\n";
+ PVE::ReplicationState::record_job_start($jobcfg, $state, $start_time, $iteration);
- my $logfunc_wrapper = sub {
- my ($msg) = @_;
+ my $t0 = [gettimeofday];
- my $ctime = get_log_time();
- print $logfd "$ctime $jobid: $msg\n";
- $logfunc->("$ctime $jobid: $msg") if $logfunc;
- };
+ mkdir $PVE::ReplicationState::replicate_logdir;
+ my $logfile = PVE::ReplicationState::job_logfile_name($jobid);
+ open(my $logfd, '>', $logfile) ||
+ die "unable to open replication log '$logfile' - $!\n";
- $logfunc_wrapper->("start replication job");
+ my $logfunc_wrapper = sub {
+ my ($msg) = @_;
- eval {
- $volumes = replicate($guest_class, $jobcfg, $state, $start_time, $logfunc_wrapper);
- };
- my $err = $@;
-
- if ($err) {
- chomp $err;
- $logfunc_wrapper->("end replication job with error: $err");
- } else {
- $logfunc_wrapper->("end replication job");
+ my $ctime = get_log_time();
+ print $logfd "$ctime $jobid: $msg\n";
+ if ($logfunc) {
+ if ($verbose) {
+ $logfunc->("$ctime $jobid: $msg");
+ } else {
+ $logfunc->($msg);
+ }
}
+ };
- PVE::ReplicationState::record_job_end($jobcfg, $state, $start_time, tv_interval($t0), $err);
+ $logfunc_wrapper->("start replication job");
- close($logfd);
+ eval {
+ $volumes = replicate($guest_class, $jobcfg, $state, $start_time, $logfunc_wrapper);
};
- if (my $err = $@) {
- warn "$jobid: got unexpected replication job error - $err";
+ my $err = $@;
+
+ if ($err) {
+ my $msg = "end replication job with error: $err";
+ chomp $msg;
+ $logfunc_wrapper->($msg);
+ } else {
+ $logfunc_wrapper->("end replication job");
}
+ PVE::ReplicationState::record_job_end($jobcfg, $state, $start_time, tv_interval($t0), $err);
+
+ close($logfd);
+
+ die $err if $err;
+
return $volumes;
};
sub run_replication {
- my ($guest_class, $jobcfg, $iteration, $start_time, $logfunc, $noerr) = @_;
+ my ($guest_class, $jobcfg, $iteration, $start_time, $logfunc, $verbose) = @_;
my $volumes;
- eval {
- my $timeout = 2; # do not wait too long - we repeat periodically anyways
- $volumes = PVE::GuestHelpers::guest_migration_lock(
- $jobcfg->{guest}, $timeout, $run_replication_nolock,
- $guest_class, $jobcfg, $iteration, $start_time, $logfunc);
- };
- if (my $err = $@) {
- return undef if $noerr;
- die $err;
- }
+ my $timeout = 2; # do not wait too long - we repeat periodically anyways
+ $volumes = PVE::GuestHelpers::guest_migration_lock(
+ $jobcfg->{guest}, $timeout, $run_replication_nolock,
+ $guest_class, $jobcfg, $iteration, $start_time, $logfunc, $verbose);
+
return $volumes;
}