|
Lines 56-191
use Try::Tiny;
Link Here
|
| 56 |
use Pod::Usage; |
56 |
use Pod::Usage; |
| 57 |
use Getopt::Long; |
57 |
use Getopt::Long; |
| 58 |
use List::MoreUtils qw( natatime ); |
58 |
use List::MoreUtils qw( natatime ); |
| 59 |
use Time::HiRes; |
|
|
| 60 |
|
59 |
|
| 61 |
use C4::Context; |
60 |
use C4::Context; |
| 62 |
use Koha::Logger; |
|
|
| 63 |
use Koha::BackgroundJobs; |
61 |
use Koha::BackgroundJobs; |
|
|
62 |
use Koha::Logger; |
| 64 |
use Koha::SearchEngine; |
63 |
use Koha::SearchEngine; |
| 65 |
use Koha::SearchEngine::Indexer; |
64 |
use Koha::SearchEngine::Indexer; |
| 66 |
|
65 |
|
| 67 |
|
66 |
my ( $max_processes, $help, $batch_size ); |
| 68 |
my ( $help, $batch_size ); |
|
|
| 69 |
|
| 70 |
my $not_found_retries = {}; |
| 71 |
my $max_retries = $ENV{MAX_RETRIES} || 10; |
| 72 |
|
| 73 |
GetOptions( |
67 |
GetOptions( |
| 74 |
'h|help' => \$help, |
68 |
'm|max-processes=i' => \$max_processes, |
| 75 |
'b|batch_size=s' => \$batch_size |
69 |
'h|help' => \$help, |
|
|
70 |
'b|batch_size=s' => \$batch_size |
| 76 |
) || pod2usage(1); |
71 |
) || pod2usage(1); |
| 77 |
|
|
|
| 78 |
pod2usage(0) if $help; |
72 |
pod2usage(0) if $help; |
| 79 |
|
|
|
| 80 |
$batch_size //= 10; |
73 |
$batch_size //= 10; |
| 81 |
|
74 |
|
| 82 |
warn "Not using Elasticsearch" unless C4::Context->preference('SearchEngine') eq 'Elasticsearch'; |
75 |
warn "Not using Elasticsearch" unless C4::Context->preference('SearchEngine') eq 'Elasticsearch'; |
| 83 |
|
|
|
| 84 |
my $logger = Koha::Logger->get({ interface => 'worker' }); |
| 85 |
|
| 86 |
my $conn; |
| 87 |
try { |
| 88 |
$conn = Koha::BackgroundJob->connect; |
| 89 |
} catch { |
| 90 |
warn sprintf "Cannot connect to the message broker, the jobs will be processed anyway (%s)", $_; |
| 91 |
}; |
| 92 |
|
| 93 |
if ( $conn ) { |
| 94 |
# FIXME cf note in Koha::BackgroundJob about $namespace |
| 95 |
my $namespace = C4::Context->config('memcached_namespace'); |
| 96 |
$conn->subscribe( |
| 97 |
{ |
| 98 |
destination => sprintf( "/queue/%s-%s", $namespace, 'elastic_index' ), |
| 99 |
ack => 'client', |
| 100 |
'prefetch-count' => 1, |
| 101 |
} |
| 102 |
); |
| 103 |
} |
| 104 |
my $biblio_indexer = Koha::SearchEngine::Indexer->new( { index => $Koha::SearchEngine::BIBLIOS_INDEX } ); |
76 |
my $biblio_indexer = Koha::SearchEngine::Indexer->new( { index => $Koha::SearchEngine::BIBLIOS_INDEX } ); |
| 105 |
my $auth_indexer = Koha::SearchEngine::Indexer->new( { index => $Koha::SearchEngine::AUTHORITIES_INDEX } ); |
77 |
my $auth_indexer = Koha::SearchEngine::Indexer->new( { index => $Koha::SearchEngine::AUTHORITIES_INDEX } ); |
| 106 |
my $config = $biblio_indexer->get_elasticsearch_params; |
78 |
my $config = $biblio_indexer->get_elasticsearch_params; |
| 107 |
my $at_a_time = $config->{chunk_size} // 5000; |
79 |
my $at_a_time = $config->{chunk_size} // 5000; |
| 108 |
|
80 |
my $lastseen = time; |
| 109 |
my @jobs = (); |
81 |
|
| 110 |
|
82 |
Koha::BackgroundWorker->run( |
| 111 |
while (1) { |
83 |
{ |
| 112 |
|
84 |
max_processes => $max_processes // $ENV{MAX_PROCESSES}, |
| 113 |
if ( $conn ) { |
85 |
mq_timeout => $ENV{MQ_TIMEOUT}, |
| 114 |
my $frame = $conn->receive_frame; |
86 |
max_retries => $ENV{MAX_RETRIES}, |
| 115 |
if ( !defined $frame ) { |
87 |
callback => \&commit, |
| 116 |
# maybe log connection problems |
88 |
queues => [ 'elastic_index' ], |
| 117 |
next; # will reconnect automatically |
|
|
| 118 |
} |
| 119 |
|
| 120 |
my $args = try { |
| 121 |
my $body = $frame->body; |
| 122 |
decode_json($body); # TODO Should this be from_json? Check utf8 flag. |
| 123 |
} catch { |
| 124 |
Koha::Logger->get({ interface => 'worker' })->warn(sprintf "Frame not processed - %s", $_); |
| 125 |
return; |
| 126 |
}; |
| 127 |
|
| 128 |
unless ( $args ) { |
| 129 |
Koha::Logger->get({ interface => 'worker' })->warn(sprintf "Frame does not have correct args, ignoring it"); |
| 130 |
$conn->nack( { frame => $frame, requeue => 'false' } ); |
| 131 |
next; |
| 132 |
} |
| 133 |
|
| 134 |
my $job = Koha::BackgroundJobs->find( $args->{job_id} ); |
| 135 |
|
| 136 |
if ( $job && $job->status ne 'new' ) { |
| 137 |
Koha::Logger->get( { interface => 'worker' } ) |
| 138 |
->warn( sprintf "Job %s has wrong status %s", $args->{job_id}, $job->status ); |
| 139 |
|
| 140 |
# nack without requeue, we do not want to process this frame again |
| 141 |
$conn->nack( { frame => $frame, requeue => 'false' } ); |
| 142 |
next; |
| 143 |
} |
| 144 |
|
| 145 |
unless ($job) { |
| 146 |
$not_found_retries->{ $args->{job_id} } //= 0; |
| 147 |
if ( ++$not_found_retries->{ $args->{job_id} } >= $max_retries ) { |
| 148 |
Koha::Logger->get( { interface => 'worker' } ) |
| 149 |
->warn( sprintf "Job %s not found, no more retry", $args->{job_id} ); |
| 150 |
|
| 151 |
# nack without requeue, we do not want to process this frame again |
| 152 |
$conn->nack( { frame => $frame, requeue => 'false' } ); |
| 153 |
next; |
| 154 |
} |
| 155 |
|
| 156 |
Koha::Logger->get( { interface => 'worker' } ) |
| 157 |
->debug( sprintf "Job %s not found, will retry later", $args->{job_id} ); |
| 158 |
|
| 159 |
# nack to force requeue |
| 160 |
$conn->nack( { frame => $frame, requeue => 'true' } ); |
| 161 |
Time::HiRes::sleep(0.5); |
| 162 |
next; |
| 163 |
} |
| 164 |
$conn->ack( { frame => $frame } ); |
| 165 |
|
| 166 |
push @jobs, $job; |
| 167 |
if ( @jobs >= $batch_size || !$conn->can_read( { timeout => '0.1' } ) ) { |
| 168 |
commit(@jobs); |
| 169 |
@jobs = (); |
| 170 |
} |
| 171 |
|
| 172 |
} else { |
| 173 |
@jobs = Koha::BackgroundJobs->search( |
| 174 |
{ status => 'new', queue => 'elastic_index' } )->as_list; |
| 175 |
commit(@jobs); |
| 176 |
@jobs = (); |
| 177 |
sleep 10; |
| 178 |
} |
89 |
} |
| 179 |
|
90 |
); |
| 180 |
} |
|
|
| 181 |
$conn->disconnect; |
| 182 |
|
91 |
|
| 183 |
sub commit { |
92 |
sub commit { |
| 184 |
my (@jobs) = @_; |
93 |
my (@jobs) = @_; |
|
|
94 |
# Extract jobs only from passed arrayrefs |
| 95 |
@jobs = map { $_->[0] } @jobs; |
| 185 |
|
96 |
|
| 186 |
my @bib_records; |
97 |
my @bib_records; |
| 187 |
my @auth_records; |
98 |
my @auth_records; |
| 188 |
|
99 |
|
|
|
100 |
my $logger = Koha::Logger->get({ interface => 'worker' }); |
| 189 |
my $jobs = Koha::BackgroundJobs->search( { id => [ map { $_->id } @jobs ] }); |
101 |
my $jobs = Koha::BackgroundJobs->search( { id => [ map { $_->id } @jobs ] }); |
| 190 |
# Start |
102 |
# Start |
| 191 |
$jobs->update({ |
103 |
$jobs->update({ |
| 192 |
- |
|
|