-
Notifications
You must be signed in to change notification settings - Fork 0
/
parallel
executable file
·12045 lines (11355 loc) · 350 KB
/
parallel
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#!/usr/bin/env perl
# Copyright (C) 2007,2008,2009,2010,2011,2012,2013,2014,2015,2016,
# 2017,2018 Ole Tange and Free Software Foundation, Inc.
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
# General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, see <http://www.gnu.org/licenses/>
# or write to the Free Software Foundation, Inc., 51 Franklin St,
# Fifth Floor, Boston, MA 02110-1301 USA
# open3 used in Job::start
use IPC::Open3;
# &WNOHANG used in reaper
use POSIX qw(:sys_wait_h setsid ceil :errno_h);
# gensym used in Job::start
use Symbol qw(gensym);
# tempfile used in Job::start
use File::Temp qw(tempfile tempdir);
# mkpath used in openresultsfile
use File::Path;
# GetOptions used in get_options_from_array
use Getopt::Long;
# Used to ensure code quality
use strict;
use File::Basename;
save_stdin_stdout_stderr();
save_original_signal_handler();
parse_options();
::debug("init", "Open file descriptors: ", join(" ",keys %Global::fd), "\n");
my $number_of_args;
if($Global::max_number_of_args) {
$number_of_args = $Global::max_number_of_args;
} elsif ($opt::X or $opt::m or $opt::xargs) {
$number_of_args = undef;
} else {
$number_of_args = 1;
}
my @command = @ARGV;
my @input_source_fh;
if($opt::pipepart) {
if($opt::tee) {
@input_source_fh = map { open_or_exit($_) } @opt::a;
# Remove the first: It will be the file piped.
shift @input_source_fh;
if(not @input_source_fh and not $opt::pipe) {
@input_source_fh = (*STDIN);
}
} else {
# -a is used for data - not for command line args
@input_source_fh = map { open_or_exit($_) } "/dev/null";
}
} else {
@input_source_fh = map { open_or_exit($_) } @opt::a;
if(not @input_source_fh and not $opt::pipe) {
@input_source_fh = (*STDIN);
}
}
if($opt::sqlmaster) {
# Create SQL table to hold joblog + output
$Global::sql->create_table($#input_source_fh+1);
if($opt::sqlworker) {
# Start a real --sqlworker in the background later
$Global::start_sqlworker = 1;
$opt::sqlworker = undef;
}
}
if($opt::skip_first_line) {
# Skip the first line for the first file handle
my $fh = $input_source_fh[0];
<$fh>;
}
set_input_source_header();
if($opt::filter_hosts and (@opt::sshlogin or @opt::sshloginfile)) {
# Parallel check all hosts are up. Remove hosts that are down
filter_hosts();
}
if($opt::nonall or $opt::onall) {
onall(\@input_source_fh,@command);
wait_and_exit(min(undef_as_zero($Global::exitstatus),254));
}
$Global::JobQueue = JobQueue->new(
\@command,\@input_source_fh,$Global::ContextReplace,
$number_of_args,\@Global::transfer_files,\@Global::ret_files);
if($opt::pipepart) {
pipepart_setup();
} elsif($opt::pipe and $opt::tee) {
pipe_tee_setup();
}
if($opt::eta or $opt::bar or $opt::shuf or $Global::halt_pct) {
# Count the number of jobs or shuffle all jobs
# before starting any.
# Must be done after ungetting any --pipepart jobs.
$Global::JobQueue->total_jobs();
}
# Compute $Global::max_jobs_running
# Must be done after ungetting any --pipepart jobs.
max_jobs_running();
init_run_jobs();
my $sem;
if($Global::semaphore) {
$sem = acquire_semaphore();
}
$SIG{TERM} = \&start_no_new_jobs;
start_more_jobs();
if($opt::tee) {
# All jobs must be running in parallel for --tee
$Global::start_no_new_jobs = 1;
} elsif($opt::pipe and not $opt::pipepart) {
spreadstdin();
}
::debug("init", "Start draining\n");
drain_job_queue();
::debug("init", "Done draining\n");
reaper();
::debug("init", "Done reaping\n");
if($Global::semaphore) {
$sem->release();
}
cleanup();
::debug("init", "Halt\n");
halt();
sub set_input_source_header {
if($opt::header and not $opt::pipe) {
# split with colsep or \t
# $header force $colsep = \t if undef?
my $delimiter = defined $opt::colsep ? $opt::colsep : "\t";
# regexp for {=
my $left = "\Q$Global::parensleft\E";
my $l = $Global::parensleft;
# regexp for =}
my $right = "\Q$Global::parensright\E";
my $r = $Global::parensright;
my $id = 1;
for my $fh (@input_source_fh) {
my $line = <$fh>;
chomp($line);
::debug("init", "Delimiter: '$delimiter'");
for my $s (split /$delimiter/o, $line) {
::debug("init", "Colname: '$s'");
# Replace {colname} with {2}
for(@command,@Global::ret_files,@Global::transfer_files,
$opt::tagstring, $opt::workdir, $opt::results, $opt::retries) {
# Skip if undefined
$_ or next;
s:\{$s(|/|//|\.|/\.)\}:\{$id$1\}:g;
# {=header1 ... =} => {=1 ... =}
s:$left $s (.*?) $right:$l$id$1$r:gx;
}
$Global::input_source_header{$id} = $s;
$id++;
}
}
} else {
my $id = 1;
for my $fh (@input_source_fh) {
$Global::input_source_header{$id} = $id;
$id++;
}
}
}
sub max_jobs_running {
# Compute $Global::max_jobs_running as the max number of jobs
# running on each sshlogin.
# Returns:
# $Global::max_jobs_running
if(not $Global::max_jobs_running) {
for my $sshlogin (values %Global::host) {
$sshlogin->max_jobs_running();
}
}
return $Global::max_jobs_running;
}
sub halt {
# Compute exit value,
# wait for children to complete
# and exit
if($opt::halt and $Global::halt_when ne "never") {
if(not defined $Global::halt_exitstatus) {
if($Global::halt_pct) {
$Global::halt_exitstatus =
::ceil($Global::total_failed / $Global::total_started * 100);
} elsif($Global::halt_count) {
$Global::halt_exitstatus =
::min(undef_as_zero($Global::total_failed),101);
}
}
wait_and_exit($Global::halt_exitstatus);
} else {
wait_and_exit(min(undef_as_zero($Global::exitstatus),101));
}
}
sub __PIPE_MODE__ {}
sub pipepart_setup {
# Compute the blocksize
# Generate the commands to extract the blocks
# Push the commands on queue
# Changes:
# @Global::cat_prepends
# $Global::JobQueue
if($opt::tee) {
# Prepend each command with
# < file
my $cat_string = "< ".::shell_quote_scalar($opt::a[0]);
for(1..$Global::JobQueue->total_jobs()) {
push @Global::cat_appends, $cat_string;
push @Global::cat_prepends, "";
}
} else {
if(not $opt::blocksize) {
# --blocksize with 10 jobs per jobslot
$opt::blocksize = -10;
}
if($opt::roundrobin) {
# --blocksize with 1 job per jobslot
$opt::blocksize = -1;
}
if($opt::blocksize < 0) {
my $size = 0;
# Compute size of -a
for(@opt::a) {
if(-f $_) {
$size += -s $_;
} elsif(-b $_) {
$size += size_of_block_dev($_);
} else {
::error("$_ is neither a file nor a block device");
wait_and_exit(255);
}
}
# Run in total $job_slots*(- $blocksize) jobs
# Set --blocksize = size / no of proc / (- $blocksize)
$Global::dummy_jobs = 1;
$Global::blocksize = 1 +
int($size / max_jobs_running() / -$opt::blocksize);
}
@Global::cat_prepends = map { pipe_part_files($_) } @opt::a;
# Unget the empty arg as many times as there are parts
$Global::JobQueue->{'commandlinequeue'}{'arg_queue'}->unget(
map { [Arg->new("\0noarg")] } @Global::cat_prepends
);
}
}
sub pipe_tee_setup {
# Create temporary fifos
# Run 'tee fifo1 fifo2 fifo3 ... fifoN' in the background
# This will spread the input to fifos
# Generate commands that reads from fifo1..N:
# cat fifo | user_command
# Changes:
# @Global::cat_prepends
my @fifos;
for(1..$Global::JobQueue->total_jobs()) {
push @fifos, tmpfifo();
}
# cat foo | tee fifo1 fifo2 fifo3 fifo4 fifo5 > /dev/null
if(not fork()){
# Let tee inherit our stdin
# and redirect stdout to null
open STDOUT, ">","/dev/null";
exec "tee",@fifos;
}
# For each fifo
# (rm fifo1; grep 1) < fifo1
# (rm fifo2; grep 2) < fifo2
# (rm fifo3; grep 3) < fifo3
# Remove the tmpfifo as soon as it is open
@Global::cat_prepends = map { "(rm $_;" } @fifos;
@Global::cat_appends = map { ") < $_" } @fifos;
}
sub pipe_part_files {
# Given the bigfile
# find header and split positions
# make commands that 'cat's the partial file
# Input:
# $file = the file to read
# Returns:
# @commands that will cat_partial each part
my ($file) = @_;
my $buf = "";
if(not -f $file and not -b $file) {
::error("$file is not a seekable file.");
::wait_and_exit(255);
}
my $header = find_header(\$buf,open_or_exit($file));
# find positions
my @pos = find_split_positions($file,$Global::blocksize,length $header);
# Make @cat_prepends
my @cat_prepends = ();
for(my $i=0; $i<$#pos; $i++) {
push(@cat_prepends,
cat_partial($file, 0, length($header), $pos[$i], $pos[$i+1]));
}
return @cat_prepends;
}
sub find_header {
# Compute the header based on $opt::header
# Input:
# $buf_ref = reference to read-in buffer
# $fh = filehandle to read from
# Uses:
# $opt::header
# $Global::blocksize
# Returns:
# $header string
my ($buf_ref, $fh) = @_;
my $header = "";
if($opt::header) {
if($opt::header eq ":") { $opt::header = "(.*\n)"; }
# Number = number of lines
$opt::header =~ s/^(\d+)$/"(.*\n)"x$1/e;
while(read($fh,substr($$buf_ref,length $$buf_ref,0),$Global::blocksize)) {
if($$buf_ref=~s/^($opt::header)//) {
$header = $1;
last;
}
}
}
return $header;
}
sub find_split_positions {
# Find positions in bigfile where recend is followed by recstart
# Input:
# $file = the file to read
# $block = (minimal) --block-size of each chunk
# $headerlen = length of header to be skipped
# Uses:
# $opt::recstart
# $opt::recend
# Returns:
# @positions of block start/end
my($file, $block, $headerlen) = @_;
my $size = -s $file;
if(-b $file) {
# $file is a blockdevice
$size = size_of_block_dev($file);
}
$block = int $block;
# The optimal dd blocksize for mint, redhat, solaris, openbsd = 2^17..2^20
# The optimal dd blocksize for freebsd = 2^15..2^17
my $dd_block_size = 131072; # 2^17
my @pos;
my ($recstart,$recend) = recstartrecend();
my $recendrecstart = $recend.$recstart;
my $fh = ::open_or_exit($file);
push(@pos,$headerlen);
for(my $pos = $block+$headerlen; $pos < $size; $pos += $block) {
my $buf;
if($recendrecstart eq "") {
# records ends anywhere
push(@pos,$pos);
} else {
# Seek the the block start
seek($fh, $pos, 0) || die;
while(read($fh,substr($buf,length $buf,0),$dd_block_size)) {
if($opt::regexp) {
# If match /$recend$recstart/ => Record position
if($buf =~ /^(.*$recend)$recstart/os) {
# Start looking for next record _after_ this match
$pos += length($1);
push(@pos,$pos);
last;
}
} else {
# If match $recend$recstart => Record position
# TODO optimize to only look at the appended
# $dd_block_size + len $recendrecstart
# TODO increase $dd_block_size to optimize for longer records
my $i = index64(\$buf,$recendrecstart);
if($i != -1) {
# Start looking for next record _after_ this match
$pos += $i + length($recend);
push(@pos,$pos);
last;
}
}
}
}
}
if($pos[$#pos] != $size) {
# Last splitpoint was not at end of the file: add it
push(@pos,$size);
}
close $fh;
return @pos;
}
sub cat_partial {
# Efficient command to copy from byte X to byte Y
# Input:
# $file = the file to read
# ($start, $end, [$start2, $end2, ...]) = start byte, end byte
# Returns:
# Efficient command to copy $start..$end, $start2..$end2, ... to stdout
my($file, @start_end) = @_;
my($start, $i);
# Convert (start,end) to (start,len)
my @start_len = map {
if(++$i % 2) { $start = $_; } else { $_-$start }
} @start_end;
# This can read 7 GB/s using a single core
my $script = spacefree
(0,
q{
while(@ARGV) {
sysseek(STDIN,shift,0) || die;
$left = shift;
while($read =
sysread(STDIN,$buf, $left > 131072 ? 131072 : $left)){
$left -= $read;
syswrite(STDOUT,$buf);
}
}
});
return "<". shell_quote_scalar($file) .
" perl -e '$script' @start_len |";
}
sub spreadstdin {
# read a record
# Spawn a job and print the record to it.
# Uses:
# $Global::blocksize
# STDIN
# $opt::r
# $Global::max_lines
# $Global::max_number_of_args
# $opt::regexp
# $Global::start_no_new_jobs
# $opt::roundrobin
# %Global::running
# Returns: N/A
if($opt::tee) {
# Spawn all jobs
# read a record
# Write record to all jobs
if(not $Global::JobQueue->empty()) {
::error("--tee requres --jobs to be higher. Try --jobs 0.");
}
}
my $buf = "";
my ($recstart,$recend) = recstartrecend();
my $recendrecstart = $recend.$recstart;
my $chunk_number = 1;
my $one_time_through;
my $two_gb = 2**31-1;
my $blocksize = $Global::blocksize;
my $in = *STDIN;
my $header = find_header(\$buf,$in);
while(1) {
my $anything_written = 0;
my $buflen = length $buf;
my $readsize = ($buflen < $blocksize) ? $blocksize-$buflen : $blocksize;
# If $buf < $blocksize, append so it is $blocksize long after reading.
# Otherwise append a full $blocksize
if(not read($in,substr($buf,$buflen,0),$readsize)) {
# End-of-file
$chunk_number != 1 and last;
# Force the while-loop once if everything was read by header reading
$one_time_through++ and last;
}
if($opt::r) {
# Remove empty lines
$buf =~ s/^\s*\n//gm;
if(length $buf == 0) {
next;
}
}
if($Global::max_lines and not $Global::max_number_of_args) {
# Read n-line records
my $n_lines = $buf =~ tr/\n/\n/;
my $last_newline_pos = rindex64(\$buf,"\n");
# Go backwards until there are full n-line records
while($n_lines % $Global::max_lines) {
$n_lines--;
$last_newline_pos = rindex64(\$buf,"\n",$last_newline_pos-1);
}
# Chop at $last_newline_pos as that is where n-line record ends
$anything_written +=
write_record_to_pipe($chunk_number++,\$header,\$buf,
$recstart,$recend,$last_newline_pos+1);
shorten(\$buf,$last_newline_pos+1);
} elsif($opt::regexp) {
if($Global::max_number_of_args) {
# -N => (start..*?end){n}
# -L -N => (start..*?end){n*l}
my $read_n_lines = -1+
$Global::max_number_of_args * ($Global::max_lines || 1);
# (?!negative lookahead) is needed to avoid backtracking
# See: https://unix.stackexchange.com/questions/439356/
while($buf =~
/# From start up till recend
^((?:(?!$recend$recstart).)*?$recend
# Then n-1 times recstart.*recend
(?:$recstart(?:(?!$recend$recstart).)*?$recend){$read_n_lines})
# Followed by recstart
(?=$recstart)/osx) {
$anything_written +=
write_record_to_pipe($chunk_number++,\$header,\$buf,
$recstart,$recend,length $1);
shorten(\$buf,length $1);
}
} else {
eof($in) and last;
# Find the last recend-recstart in $buf
if($buf =~ /^(.*$recend)$recstart.*?$/os) {
$anything_written +=
write_record_to_pipe($chunk_number++,\$header,\$buf,
$recstart,$recend,length $1);
shorten(\$buf,length $1);
}
}
} elsif($opt::csv) {
# Read a full CSV record
# even number of " + end of line
my $last_newline_pos = length $buf;
do {
# find last EOL
$last_newline_pos = rindex64(\$buf,"\n",$last_newline_pos-1);
# While uneven "
} while((substr($buf,0,$last_newline_pos) =~ y/"/"/)%2
and $last_newline_pos >= 0);
# Chop at $last_newline_pos as that is where CSV record ends
$anything_written +=
write_record_to_pipe($chunk_number++,\$header,\$buf,
$recstart,$recend,$last_newline_pos+1);
shorten(\$buf,$last_newline_pos+1);
} else {
if($Global::max_number_of_args) {
# -N => (start..*?end){n}
my $i = 0;
my $read_n_lines =
$Global::max_number_of_args * ($Global::max_lines || 1);
while(($i = nindex(\$buf,$recendrecstart,$read_n_lines)) != -1
and
length $buf) {
$i += length $recend; # find the actual splitting location
$anything_written +=
write_record_to_pipe($chunk_number++,\$header,\$buf,
$recstart,$recend,$i);
shorten(\$buf,$i);
}
} else {
eof($in) and last;
# Find the last recend+recstart in $buf
my $i = rindex64(\$buf,$recendrecstart);
if($i != -1) {
$i += length $recend; # find the actual splitting location
$anything_written +=
write_record_to_pipe($chunk_number++,\$header,\$buf,
$recstart,$recend,$i);
shorten(\$buf,$i);
}
}
}
if(not $anything_written
and not eof($in)
and not $Global::no_autoexpand_block) {
# Nothing was written - maybe the block size < record size?
# Increase blocksize exponentially up to 2GB-1 (2GB causes problems)
if($blocksize < $two_gb) {
my $old_blocksize = $blocksize;
$blocksize = ::min(ceil($blocksize * 1.3 + 1), $two_gb);
::warning("A record was longer than $old_blocksize. " .
"Increasing to --blocksize $blocksize.");
}
}
}
::debug("init", "Done reading input\n");
# If there is anything left in the buffer write it
write_record_to_pipe($chunk_number++, \$header, \$buf, $recstart,
$recend, length $buf);
if($opt::retries) {
$Global::no_more_input = 1;
# We need to start no more jobs: At most we need to retry some
# of the already running.
my @running = values %Global::running;
# Stop any virgins.
for my $job (@running) {
if(defined $job and $job->virgin()) {
close $job->fh(0,"w");
}
}
# Wait for running jobs to be done
my $sleep =1;
while($Global::total_running > 0) {
$sleep = ::reap_usleep($sleep);
}
}
$Global::start_no_new_jobs ||= 1;
if($opt::roundrobin) {
# Flush blocks to roundrobin procs
my $sleep = 1;
while(%Global::running) {
my $something_written = 0;
for my $job (values %Global::running) {
if($job->block_length()) {
$something_written += $job->non_blocking_write();
} else {
close $job->fh(0,"w");
}
}
if($something_written) {
$sleep = $sleep/2+0.001;
}
$sleep = ::reap_usleep($sleep);
}
}
}
sub recstartrecend {
# Uses:
# $opt::recstart
# $opt::recend
# Returns:
# $recstart,$recend with default values and regexp conversion
my($recstart,$recend);
if(defined($opt::recstart) and defined($opt::recend)) {
# If both --recstart and --recend is given then both must match
$recstart = $opt::recstart;
$recend = $opt::recend;
} elsif(defined($opt::recstart)) {
# If --recstart is given it must match start of record
$recstart = $opt::recstart;
$recend = "";
} elsif(defined($opt::recend)) {
# If --recend is given then it must match end of record
$recstart = "";
$recend = $opt::recend;
if($opt::regexp and $recend eq '') {
# --regexp --recend ''
$recend = '.';
}
}
if($opt::regexp) {
# If $recstart/$recend contains '|' this should only apply to the regexp
$recstart = "(?:".$recstart.")";
$recend = "(?:".$recend.")";
} else {
# $recstart/$recend = printf strings (\n)
$recstart =~ s/\\([0rnt\'\"\\])/"qq|\\$1|"/gee;
$recend =~ s/\\([0rnt\'\"\\])/"qq|\\$1|"/gee;
}
return ($recstart,$recend);
}
sub nindex {
# See if string is in buffer N times
# Returns:
# the position where the Nth copy is found
my ($buf_ref, $str, $n) = @_;
my $i = 0;
for(1..$n) {
$i = index64($buf_ref,$str,$i+1);
if($i == -1) { last }
}
return $i;
}
{
my @robin_queue;
my $sleep = 1;
sub round_robin_write {
# Input:
# $header_ref = ref to $header string
# $block_ref = ref to $block to be written
# $recstart = record start string
# $recend = record end string
# $endpos = end position of $block
# Uses:
# %Global::running
# Returns:
# $something_written = amount of bytes written
my ($header_ref,$buffer_ref,$recstart,$recend,$endpos) = @_;
my $written = 0;
my $block_passed = 0;
while(not $block_passed) {
# Continue flushing existing buffers
# until one is empty and a new block is passed
if(@robin_queue) {
# Rotate queue once so new blocks get a fair chance
# to be given to another block
push @robin_queue, shift @robin_queue;
} else {
# Make a queue to spread the blocks evenly
push @robin_queue, (sort { $a->seq() <=> $b->seq() }
values %Global::running);
}
if($opt::keeporder) {
for my $job (@robin_queue) {
if($job->block_length() > 0) {
$written += $job->non_blocking_write();
} else {
$job->set_block($header_ref,$buffer_ref,$endpos,$recstart,$recend);
$block_passed = 1;
$job->set_virgin(0);
$written += $job->non_blocking_write();
last;
}
}
} else {
do {
$written = 0;
for my $job (@robin_queue) {
if($job->block_length() > 0) {
$written += $job->non_blocking_write();
} else {
$job->set_block($header_ref,$buffer_ref,
$endpos,$recstart,$recend);
$block_passed = 1;
$job->set_virgin(0);
$written += $job->non_blocking_write();
last;
}
}
if($written) {
$sleep = $sleep/1.5+0.001;
}
} while($written and not $block_passed);
}
$sleep = ::reap_usleep($sleep);
}
return $written;
}
}
sub index64 {
# Do index on strings > 2GB.
# index in Perl < v5.22 does not work for > 2GB
# Input:
# as index except STR which must be passed as a reference
# Output:
# as index
my $ref = shift;
my $match = shift;
my $pos = shift || 0;
my $block_size = 2**31-1;
my $strlen = length($$ref);
# No point in doing extra work if we don't need to.
if($strlen < $block_size or $] > 5.022) {
return index($$ref, $match, $pos);
}
my $matchlen = length($match);
my $ret;
my $offset = $pos;
while($offset < $strlen) {
$ret = index(
substr($$ref, $offset, $block_size),
$match, $pos-$offset);
if($ret != -1) {
return $ret + $offset;
}
$offset += ($block_size - $matchlen - 1);
}
return -1;
}
sub rindex64 {
# Do rindex on strings > 2GB.
# rindex in Perl < v5.22 does not work for > 2GB
# Input:
# as rindex except STR which must be passed as a reference
# Output:
# as rindex
my $ref = shift;
my $match = shift;
my $pos = shift;
my $block_size = 2**31-1;
my $strlen = length($$ref);
# Default: search from end
$pos = defined $pos ? $pos : $strlen;
# No point in doing extra work if we don't need to.
if($strlen < $block_size) {
return rindex($$ref, $match, $pos);
}
my $matchlen = length($match);
my $ret;
my $offset = $pos - $block_size + $matchlen;
if($offset < 0) {
# The offset is less than a $block_size
# Set the $offset to 0 and
# Adjust block_size accordingly
$block_size = $block_size + $offset;
$offset = 0;
}
while($offset >= 0) {
$ret = rindex(
substr($$ref, $offset, $block_size),
$match);
if($ret != -1) {
return $ret + $offset;
}
$offset -= ($block_size - $matchlen - 1);
}
return -1;
}
sub shorten {
# Do: substr($buf,0,$i) = "";
# Some Perl versions do not support $i > 2GB, so do this in 2GB chunks
# Input:
# $buf_ref = \$buf
# $i = position to shorten to
# Returns: N/A
my ($buf_ref, $i) = @_;
my $two_gb = 2**31-1;
while($i > $two_gb) {
substr($$buf_ref,0,$two_gb) = "";
$i -= $two_gb;
}
substr($$buf_ref,0,$i) = "";
}
sub write_record_to_pipe {
# Fork then
# Write record from pos 0 .. $endpos to pipe
# Input:
# $chunk_number = sequence number - to see if already run
# $header_ref = reference to header string to prepend
# $buffer_ref = reference to record to write
# $recstart = start string of record
# $recend = end string of record
# $endpos = position in $buffer_ref where record ends
# Uses:
# $Global::job_already_run
# $opt::roundrobin
# @Global::virgin_jobs
# Returns:
# Number of chunks written (0 or 1)
my ($chunk_number,$header_ref,$buffer_ref,$recstart,$recend,$endpos) = @_;
if($endpos == 0) { return 0; }
if(vec($Global::job_already_run,$chunk_number,1)) { return 1; }
if($opt::roundrobin) {
# Write the block to one of the already running jobs
return round_robin_write($header_ref,$buffer_ref,$recstart,$recend,$endpos);
}
# If no virgin found, backoff
my $sleep = 0.0001; # 0.01 ms - better performance on highend
while(not @Global::virgin_jobs) {
::debug("pipe", "No virgin jobs");
$sleep = ::reap_usleep($sleep);
# Jobs may not be started because of loadavg
# or too little time between each ssh login
# or retrying failed jobs.
start_more_jobs();
}
my $job = shift @Global::virgin_jobs;
# Job is no longer virgin
$job->set_virgin(0);
if($opt::retries) {
# Copy $buffer[0..$endpos] to $job->{'block'}
# Remove rec_sep
# Run $job->add_transfersize
$job->set_block($header_ref,$buffer_ref,$endpos,$recstart,$recend);
if(fork()) {
# Skip
} else {
$job->write($job->block_ref());
close $job->fh(0,"w");
exit(0);
}
} else {
# We ignore the removed rec_sep which is technically wrong.
$job->add_transfersize($endpos + length $$header_ref);
if(fork()) {
# Skip
} else {
# Chop of at $endpos as we do not know how many rec_sep will
# be removed.
substr($$buffer_ref,$endpos,length $$buffer_ref) = "";
# Remove rec_sep
if($opt::remove_rec_sep) {
Job::remove_rec_sep($buffer_ref,$recstart,$recend);
}
$job->write($header_ref);
$job->write($buffer_ref);
close $job->fh(0,"w");
exit(0);
}
}
close $job->fh(0,"w");
return 1;
}
sub __SEM_MODE__ {}
sub acquire_semaphore {
# Acquires semaphore. If needed: spawns to the background
# Uses:
# @Global::host
# Returns:
# The semaphore to be released when jobs is complete
$Global::host{':'} = SSHLogin->new(":");
my $sem = Semaphore->new($Semaphore::name,$Global::host{':'}->max_jobs_running());
$sem->acquire();
if($Semaphore::fg) {
# skip
} else {
if(fork()) {
exit(0);
} else {
# If run in the background, the PID will change
$sem->pid_change();
}
}
return $sem;
}
sub __PARSE_OPTIONS__ {}
sub options_hash {
# Returns:
# %hash = the GetOptions config
return
("debug|D=s" => \$opt::D,
"xargs" => \$opt::xargs,
"m" => \$opt::m,
"X" => \$opt::X,
"v" => \@opt::v,
"sql=s" => \$opt::retired,
"sqlmaster=s" => \$opt::sqlmaster,
"sqlworker=s" => \$opt::sqlworker,
"sqlandworker=s" => \$opt::sqlandworker,
"joblog|jl=s" => \$opt::joblog,
"results|result|res=s" => \$opt::results,
"resume" => \$opt::resume,
"resume-failed|resumefailed" => \$opt::resume_failed,
"retry-failed|retryfailed" => \$opt::retry_failed,
"silent" => \$opt::silent,
"keep-order|keeporder|k" => \$opt::keeporder,
"no-keep-order|nokeeporder|nok|no-k" => \$opt::nokeeporder,
"group" => \$opt::group,
"g" => \$opt::retired,
"ungroup|u" => \$opt::ungroup,
"linebuffer|linebuffered|line-buffer|line-buffered|lb" => \$opt::linebuffer,
"tmux" => \$opt::tmux,
"tmuxpane" => \$opt::tmuxpane,
"null|0" => \$opt::null,
"quote|q" => \$opt::q,
# Replacement strings
"parens=s" => \$opt::parens,
"rpl=s" => \@opt::rpl,
"plus" => \$opt::plus,
"I=s" => \$opt::I,
"extensionreplace|er=s" => \$opt::U,
"U=s" => \$opt::retired,
"basenamereplace|bnr=s" => \$opt::basenamereplace,
"dirnamereplace|dnr=s" => \$opt::dirnamereplace,
"basenameextensionreplace|bner=s" => \$opt::basenameextensionreplace,
"seqreplace=s" => \$opt::seqreplace,
"slotreplace=s" => \$opt::slotreplace,
"jobs|j=s" => \$opt::jobs,
"delay=s" => \$opt::delay,
"sshdelay=f" => \$opt::sshdelay,
"load=s" => \$opt::load,
"noswap" => \$opt::noswap,
"max-line-length-allowed" => \$opt::max_line_length_allowed,
"number-of-cpus" => \$opt::number_of_cpus,