source: main/trunk/greenstone2/perllib/servercontrol.pm@ 30524

Last change on this file since 30524 was 30524, checked in by ak19, 8 years ago
  1. get_library_URL() called internal to servercontrol.pm when using servercontrol in activate.pl too. 2. More tidying up: removed lots of variables no longer in use.
File size: 21.6 KB
Line 
1#############################################################################
2#
3# activate.pm -- functions to get the GS library URL, ping the library URL,
4# activate and deactivate a collection.
5#
6# A component of the Greenstone digital library software
7# from the New Zealand Digital Library Project at the
8# University of Waikato, New Zealand.
9#
10# Copyright (C) 1999 New Zealand Digital Library Project
11#
12# This program is free software; you can redistribute it and/or modify
13# it under the terms of the GNU General Public License as published by
14# the Free Software Foundation; either version 2 of the License, or
15# (at your option) any later version.
16#
17# This program is distributed in the hope that it will be useful,
18# but WITHOUT ANY WARRANTY; without even the implied warranty of
19# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20# GNU General Public License for more details.
21#
22# You should have received a copy of the GNU General Public License
23# along with this program; if not, write to the Free Software
24# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
25#
26###############################################################################
27
28package servercontrol;
29
30
31use strict;
32no strict 'refs'; # allow filehandles to be variables and vice versa
33no strict 'subs'; # allow barewords (eg STDERR) as function arguments
34
35use HTTP::Response;
36use LWP::Simple qw($ua !head); # import useragent object as $ua from the full LWP to use along with LWP::Simple
37 # don't import LWP::Simple's head function by name since it can conflict with CGI:head())
38#use CGI qw(:standard); # then only CGI.pm defines a head()
39use Net::Ping;
40use URI;
41
42use printusage;
43use parse2;
44
45
46sub new
47{
48 my $class = shift(@_);
49
50 my ($qualified_collection, $site, $verbosity, $build_dir, $index_dir, $collect_dir, $library_url, $library_name) = @_;
51
52 # library_url: to be specified on the cmdline if not using a GS-included web server
53 # the GSDL_LIBRARY_URL env var is useful when running cmdline buildcol.pl in the linux package manager versions of GS3
54
55 my $self = {'build_dir' => $build_dir,
56 'index_dir' => $index_dir,
57 'collect_dir' => $collect_dir,
58 'site' => $site,
59 'qualified_collection' => $qualified_collection,
60 #'is_persistent_server' => undef,
61 'library_url' => $library_url || $ENV{'GSDL_LIBRARY_URL'} || undef, # to be specified on the cmdline if not using a GS-included web server
62 'library_name' => $library_name,
63 #'gs_mode' => "gs2",
64 'verbosity' => $verbosity || 2
65 };
66
67
68 # Do we need 'listall' support in buildcol? If so, copy code from inexport
69 # later [jmt12]
70
71 if ((defined $site) && ($site ne "")) { # GS3
72 $self->{'gs_mode'} = "gs3";
73 } else {
74 $self->{'gs_mode'} = "gs2";
75 }
76
77 return bless($self, $class);
78}
79
80## TODO: gsprintf to $self->{'out'} in these 2 print functions
81## See buildcolutils.pm new() for setting up $out
82
83sub print_task_msg {
84 my $self = shift(@_);
85 my ($task_msg, $verbosity_setting) = @_;
86
87 $verbosity_setting = $self->{'verbosity'} unless $verbosity_setting;
88 #$verbosity_setting = 1 unless defined $verbosity;
89 if($verbosity_setting >= 1) {
90 print STDERR "\n";
91 print STDERR "************************\n";
92 print STDERR "* $task_msg\n";
93 print STDERR "************************\n";
94 }
95}
96
97# Prints messages if the verbosity is right. Does not add new lines.
98sub print_msg {
99 my $self = shift(@_);
100 my ($msg, $min_verbosity, $verbosity_setting) = @_;
101
102 # only display error messages if the current
103 # verbosity setting >= the minimum verbosity level
104 # needed for that message to be displayed.
105
106 $verbosity_setting = $self->{'verbosity'} unless defined $verbosity_setting;
107 $min_verbosity = 1 unless defined $min_verbosity;
108 if($verbosity_setting >= $min_verbosity) { # by default display all 1 messages
109 print STDERR "$msg";
110 }
111}
112
113# Method to send a command to a GS2 or GS3 library_URL
114# the commands used in this script can be activate, deactivate, ping,
115# and is-persistent (is-persistent only implemented for GS2).
116sub config {
117 my $self = shift(@_);
118 my ($command, $check_message_against_regex, $expected_error_code, $silent) = @_;
119
120 my $library_url = $self->get_library_URL(); #$self->{'library_url'};
121
122
123 # Gatherer.java's configGS3Server doesn't use the site variable
124 # so we don't have to either
125
126 # for GS2, getting the HTTP status isn't enough, we need to read the output
127 # since this is what CollectionManager.config() stipulates.
128 # Using LWP::UserAgent::get($url) for this
129
130 if(!defined $library_url) {
131 return 0;
132 }
133 else {
134 $ua->timeout(5); # set LWP useragent to 5s max timeout for testing the URL
135 # Need to set this, else it takes I don't know how long to timeout
136 # http://www.perlmonks.org/?node_id=618534
137
138 # http://search.cpan.org/~gaas/libwww-perl-6.04/lib/LWP/UserAgent.pm
139 # use LWP::UserAgent's get($url) since it returns an HTTP::Response code
140
141 my $response_obj = $ua->get($library_url.$command);
142
143 # $response_obj->content stores the content and $response_obj->code the HTTP response code
144 my $response_code = $response_obj->code();
145
146 if(LWP::Simple::is_success($response_code)) {# $response_code eq RC_OK) { # LWP::Simple::is_success($response_code)
147 $self->print_msg("*** Command $library_url$command\n", 3);
148 $self->print_msg("*** HTTP Response Status: $response_code - Complete.", 3);
149
150 # check the page content is as expected
151 my $response_content = $response_obj->content;
152 my $resultstr = $response_content;
153 $resultstr =~ s@.*gs_content\"\>@@s;
154 $resultstr =~ s@</div>.*@@s;
155
156 if($response_content =~ m/$check_message_against_regex/) {
157 $self->print_msg(" Response as expected.\n", 3);
158 $self->print_msg("@@@@@@ Got result:\n$resultstr\n", 4);
159 return 1;
160 } else {
161 # if we expect the collection to be inactive, then we'd be in silent mode: if so,
162 # don't print out the "ping did not succeed" response, but print out any other messages
163
164 # So we only suppress the ping col "did not succeed" response if we're in silent mode
165 # But if any message other than ping "did not succeed" is returned, we always print it
166 if($response_content !~ m/did not succeed/ || !$silent) {
167 $self->print_msg("\n\tBUT: command $library_url$command response UNEXPECTED.\n", 3);
168 $self->print_msg("*** Got message:\n$response_content.\n", 4);
169 $self->print_msg("*** Got result:\n$resultstr\n", 3);
170 }
171 return 0; # ping on a collection may "not succeed."
172 }
173 }
174 elsif(LWP::Simple::is_error($response_code)) { # method exported by LWP::Simple, along with HTTP::Status constants
175 # check the page content is as expected
176 if(defined $expected_error_code && $response_code == $expected_error_code) {
177 $self->print_msg(" Response status $response_code as expected.\n", 3);
178 } else {
179 $self->print_msg("*** Command $library_url$command\n");
180 $self->print_msg("*** Unexpected error. HTTP Response Status: $response_code - Failed.\n");
181 }
182 return 0; # return false, since the response_code was an error, expected or not
183 }
184 else {
185 $self->print_msg("*** Command $library_url$command\n");
186 $self->print_msg("*** Unexpected error. HTTP Response Status: $response_code - Failed.\n");
187 return 0;
188 }
189 }
190}
191
192sub deactivate_collection {
193 my $self = shift(@_);
194
195 my $gs_mode = $self->{'gs_mode'};
196 my $qualified_collection = $self->{'qualified_collection'};
197
198 if($gs_mode eq "gs2") {
199 my $DEACTIVATE_COMMAND = "?a=config&cmd=release-collection&c=";
200 my $check_message_against_regex = q/configured release-collection/;
201 $self->config($DEACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
202 }
203 elsif ($gs_mode eq "gs3") {
204 my $DEACTIVATE_COMMAND = "?a=s&sa=d&st=collection&sn=";
205 my $check_message_against_regex = "collection: $qualified_collection deactivated";
206 $self->config($DEACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
207 }
208}
209
210sub activate_collection {
211 my $self = shift(@_);
212
213 my $gs_mode = $self->{'gs_mode'};
214 my $qualified_collection = $self->{'qualified_collection'};
215
216 if($gs_mode eq "gs2") {
217 my $ACTIVATE_COMMAND = "?a=config&cmd=add-collection&c=";
218 my $check_message_against_regex = q/configured add-collection/;
219 $self->config($ACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
220 }
221 elsif ($gs_mode eq "gs3") {
222 my $ACTIVATE_COMMAND = "?a=s&sa=a&st=collection&sn=";
223 my $check_message_against_regex = "collection: $qualified_collection activated";
224 $self->config($ACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
225 }
226}
227
228sub ping {
229 my $self = shift(@_);
230 my $command = shift(@_);
231 my $silent = shift(@_);
232
233 # If the GS server is not running, we *expect* to see a "500" status code.
234 # If the GS server is running, then "Ping" ... "succeeded" is expected on success.
235 # When pinging an inactive collection, it will say it did "not succeed". This is
236 # a message of interest to return.
237 my $check_responsemsg_against_regex = q/(succeeded)/;
238 my $expected_error_code = 500;
239
240 $self->print_msg("*** COMMAND WAS: |$command|***\n", 4);
241
242 return $self->config($command, $check_responsemsg_against_regex, $expected_error_code, $silent);
243}
244
245# send a pingaction to the GS library. General server-level ping.
246sub ping_library {
247 my $self = shift(@_);
248
249 my $gs_mode = $self->{'gs_mode'};
250
251 my $command = "";
252 if($gs_mode eq "gs2") {
253 $command = "?a=ping";
254 }
255 elsif ($gs_mode eq "gs3") {
256 $command = "?a=s&sa=ping";
257 }
258 return $self->ping($command);
259}
260
261
262# send a pingaction to a collection in GS library to check if it's active
263sub ping_library_collection {
264 my $self = shift(@_);
265 my $silent = shift(@_);
266
267 my $gs_mode = $self->{'gs_mode'};
268 my $qualified_collection = $self->{'qualified_collection'};
269
270 my $command = "";
271 if($gs_mode eq "gs2") {
272 $command = "?a=ping&c=$qualified_collection";
273 }
274 elsif ($gs_mode eq "gs3") {
275 $command = "?a=s&sa=ping&st=collection&sn=$qualified_collection";
276 }
277 return $self->ping($command, $silent);
278}
279
280# return true if server is persistent, by calling is-persistent on library_url
281# this is only for GS2, since the GS3 server is always persistent
282sub is_persistent {
283 my $self = shift(@_);
284
285 if($self->{'gs_mode'} eq "gs3") { # GS3 server is always persistent
286 return 1;
287 }
288
289 my $command = "?a=is-persistent";
290 my $check_responsemsg_against_regex = q/true/; # isPersistent: true versus isPersistent: false
291 return $self->config($command, $check_responsemsg_against_regex);
292}
293
294sub set_library_URL {
295 my $self = shift(@_);
296 my $library_url = shift(@_);
297 $self->{'library_url'} = $library_url;
298}
299
300sub get_library_URL {
301 my $self = shift(@_);
302
303 # For web servers that are external to a Greenstone installation,
304 # the user can pass in their web server's library URL.
305 if($self->{'library_url'}) {
306 return $self->{'library_url'};
307 }
308
309 # For web servers included with GS (like tomcat for GS3 and server.exe
310 # and apache for GS2), we work out the library URL:
311 my ($gs_mode, $lib_name); # gs_mode can be gs3 or gs2, lib_name is the custom servlet name
312 $gs_mode = $self->{'gs_mode'};
313 $lib_name = $self->{'library_name'};
314
315 # If we get here, we are dealing with a server included with GS.
316 # For GS3, we ask ant for the library URL.
317 # For GS2, we derive the URL from the llssite.cfg file.
318
319 my $url = undef;
320
321 if($gs_mode eq "gs2") {
322 my $llssite_cfg = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "llssite.cfg");
323
324 if(-f $llssite_cfg) {
325 # check llssite.cfg for line with url property
326 # for server.exe also need to use portnumber and enterlib properties
327
328 # Read in the entire contents of the file in one hit
329 if (!open (FIN, $llssite_cfg)) {
330 $self->print_msg("activate.pl::get_library_URL failed to open $llssite_cfg ($!)\n");
331 return undef;
332 }
333
334 my $contents;
335 sysread(FIN, $contents, -s FIN);
336 close(FIN);
337
338 my @lines = split(/[\n\r]+/, $contents); # split on carriage-returns and/or linefeeds
339 my $enterlib = "";
340 my $portnumber = ""; # will remain empty (implicit port 80) unless it's specifically been assigned
341
342 foreach my $line (@lines) {
343 if($line =~ m/^url=(.*)$/) {
344 $url = $1;
345 } elsif($line =~ m/^enterlib=(.*)$/) {
346 $enterlib = $1;
347 } elsif($line =~ m/^portnumber=(.*)$/) {
348 $portnumber = $1;
349 }
350 }
351
352 if(!$url) {
353 return undef;
354 }
355 elsif($url eq "URL_pending") { # library is not running
356 # do not process url=URL_pending in the file, since for server.exe
357 # this just means the Enter Library button hasn't been pressed yet
358 $url = undef;
359 }
360 else {
361 # In the case of server.exe, need to do extra work to get the proper URL
362 # But first, need to know whether we're indeed dealing with server.exe:
363
364 # compare the URL's domain to the full URL
365 # E.g. for http://localhost:8383/greenstone3/cgi-bin, the domain is localhost:8383
366 my $uri = URI->new( $url );
367 my $host = $uri->host;
368 #print STDERR "@@@@@ host: $host\n";
369 if($url =~ m/http:\/\/$host(\/)?$/) {
370 #if($url !~ m/http:\/\/$host:$portnumber(\/)?/ || $url =~ m/http:\/\/$host(\/)?$/) {
371 # (if the URL does not contain the portnumber, OR if the port is implicitly 80 and)
372 # If the domain with http:// prefix is completely the same as the URL, assume server.exe
373 # then the actual URL is the result of suffixing the port and enterlib properties in llssite.cfg
374 $url = $url.":".$portnumber.$enterlib;
375 } # else, apache web server
376
377 }
378 }
379 } elsif($gs_mode eq "gs3") {
380 # Either check build.properties for tomcat.server, tomcat.port and app.name (and default servlet name).
381 # app.name is stored in app.path by build.xml. Need to move app.name in build.properties from build.xml
382
383 # Or, run the new target get-default-servlet-url
384 # the output can look like:
385 #
386 # Buildfile: build.xml
387 # [echo] os.name: Windows Vista
388 #
389 # get-default-servlet-url:
390 # [echo] http://localhost:8383/greenstone3/library
391 # BUILD SUCCESSFUL
392 # Total time: 0 seconds
393
394 #my $output = qx/ant get-default-servlet-url/; # backtick operator, to get STDOUT (else 2>&1)
395 # see http://stackoverflow.com/questions/799968/whats-the-difference-between-perls-backticks-system-and-exec
396
397 # The get-default-servlet-url ant target can be run from anywhere by specifying the
398 # location of GS3's ant build.xml buildfile. Activate.pl can be run from anywhere for GS3
399 # GSDL3SRCHOME will be set for GS3 by gs3-setup.sh, a step that would have been necessary
400 # to run the activate.pl script in the first place
401 my $perl_command = "ant -buildfile \"$ENV{'GSDL3SRCHOME'}/build.xml\" get-default-servlet-url";
402
403 if (open(PIN, "$perl_command |")) {
404 while (defined (my $perl_output_line = <PIN>)) {
405 if($perl_output_line =~ m@http:\/\/(\S*)@) { # grab all the non-whitespace chars
406 $url="http://".$1;
407 }
408 }
409 close(PIN);
410 } else {
411 $self->print_msg("activate.pl::get_library_URL: Failed to run $perl_command to work out library URL for $gs_mode\n");
412 }
413 if(defined $lib_name) {
414 # replace the servlet_name portion of the url found, with the given library_name
415 $url =~ s@/[^/]*$@/$lib_name@;
416 }
417 }
418
419 # either the url is still undef or it is now set
420 #print STDERR "\n@@@@@ final URL:|$url|\n" if $url;
421 #print STDERR "\n@@@@@ URL still undef\n" if !$url;
422
423 $self->{'library_url'} = $url;
424 return $url;
425}
426
427
428sub do_deactivate {
429 my $self = shift(@_);
430
431 # 1. Get library URL
432
433 # For web servers that are external to a Greenstone installation,
434 # the user can pass in their web server's library URL.
435 # For web servers included with GS (like tomcat for GS3 and server.exe
436 # and apache for GS2), we work out the library URL:
437
438 # Can't do $self->{'library_url'}, because it may not yet be set
439 my $library_url = $self->get_library_URL(); # returns undef if no server is running
440
441 if(!defined $library_url) { # undef if no valid server URL
442 return; # can't do any deactivation without a running GS server
443 }
444
445 my $is_persistent_server = $self->{'is_persistent_server'};
446 my $qualified_collection = $self->{'qualified_collection'};
447
448 # CollectionManager's installCollection phase in GLI
449 # 2. Ping the library URL, and if it's a persistent server and running, release the collection
450
451 $self->print_msg("Pinging $library_url\n");
452 if ($self->ping_library()) { # server running
453
454 # server is running, so release the collection if
455 # the server is persistent and the collection is active
456
457 # don't need to work out persistency of server more than once, since the libraryURL hasn't changed
458 if (!defined $is_persistent_server) {
459 $self->print_msg("Checking if Greenstone server is persistent\n");
460 $is_persistent_server = $self->is_persistent();
461 $self->{'is_persistent_server'} = $is_persistent_server;
462 }
463
464 if ($is_persistent_server) { # only makes sense to issue activate and deactivate cmds to a persistent server
465
466 $self->print_msg("Checking if the collection $qualified_collection is already active\n");
467 my $collection_active = $self->ping_library_collection();
468
469 if ($collection_active) {
470 $self->print_msg("De-activating collection $qualified_collection\n");
471 $self->deactivate_collection();
472 }
473 else {
474 $self->print_msg("Collection is not active => No need to deactivate\n");
475 }
476 }
477 else {
478 $self->print_msg("Server is not persistent => No need to deactivate collection\n");
479 }
480 }
481 else {
482 $self->print_msg("No response to Ping => Taken to mean server is not running\n");
483 }
484
485 return $is_persistent_server;
486}
487
488sub do_activate {
489 my $self = shift @_;
490
491 my $library_url = $self->get_library_URL(); # Can't do $self->{'library_url'}; as it may not be set yet
492
493 if(!defined $library_url) { # undef if no valid server URL
494 return; # nothing to activate if there's no running server
495 }
496
497 my $is_persistent_server = $self->{'is_persistent_server'};
498 my $qualified_collection = $self->{'qualified_collection'};
499
500 $self->print_msg("Pinging $library_url\n");
501 if ($self->ping_library()) { # server running
502
503 # don't need to work out persistency of server more than once, since the libraryURL hasn't changed
504 if (!defined $is_persistent_server) {
505 $self->print_msg("Checking if Greenstone server is persistent\n");
506 $is_persistent_server = $self->is_persistent();
507 $self->{'is_persistent_server'} = $is_persistent_server;
508 }
509
510 if ($is_persistent_server) { # persistent server, so can try activating collection
511
512 $self->print_msg("Checking if the collection $qualified_collection is not already active\n");
513
514 # Since we could have deactivated the collection at this point,
515 # it is likely that it is not yet active. When pinging the collection
516 # a "ping did not succeed" message is expected, therefore tell the ping
517 # to proceed silently
518 my $silent = 1;
519 my $collection_active = $self->ping_library_collection($silent);
520
521 if (!$collection_active) {
522 $self->print_msg(" Collection is not active.\n");
523 $self->print_msg("Activating collection $qualified_collection\n");
524 $self->activate_collection();
525
526 # unless an error occurred, the collection should now be active:
527 $collection_active = $self->ping_library_collection(); # not silent if ping did not succeed
528 if(!$collection_active) {
529 $self->print_msg("ERROR: collection $qualified_collection did not get activated\n");
530 }
531 }
532 else {
533 $self->print_msg("Collection is already active => No need to activate\n");
534 }
535 }
536 else {
537 $self->print_msg("Server is not persistent => No need to activate collection\n");
538 }
539 }
540 else {
541 $self->print_msg("No response to Ping => Taken to mean server is not running\n");
542 }
543
544 return $is_persistent_server;
545}
546
547
548#########################################################
549### UNUSED METHODS - CAN BE HANDY
550
551# This method is now unused. Using ping_library instead to send the ping action to a
552# GS2/GS3 server. This method can be used more generally to test whether a URL is alive.
553# http://search.cpan.org/dist/libwww-perl/lib/LWP/Simple.pm
554# and http://www.perlmonks.org/?node_id=618534
555sub is_URL_active {
556 my $url = shift(@_); # gs3 or gs2 URL
557
558 my $status = 0;
559 if(defined $url) {
560 $ua->timeout(10); # set LWP useragent to 5s max timeout for testing the URL
561 # Need to set this, else it takes I don't know how long to timeout
562 # http://www.perlmonks.org/?node_id=618534
563
564 $status = LWP::Simple::head($url); # returns empty list of headers if it fails
565 # LWP::Simple::get($url) is more intensive, so don't need to do that
566 #print STDERR "**** $url is alive.\n" if $status;
567 }
568 return $status;
569}
570
571# Pinging seems to always return true, so this method doesn't work
572sub pingHost {
573 my $url = shift(@_); # gs3 or gs2 URL
574
575 my $status = 0;
576 if(defined $url) {
577 # Get just the domain. "http://localhost/gsdl?uq=332033495" becomes "localhost"
578 # "http://localhost/greenstone/cgi-bin/library.cgi" becomes "localhost" too
579
580 #my $host = $url;
581 #$host =~ s@^http:\/\/(www.)?@@;
582 #$host =~ s@\/.*@@;
583 #print STDERR "**** HOST: $host\n";
584
585 # More robust way
586 # http://stackoverflow.com/questions/827024/how-do-i-extract-the-domain-out-of-an-url
587 my $uri = URI->new( $url );
588 my $host = $uri->host;
589
590 # Ping the host. http://perldoc.perl.org/Net/Ping.html
591 my $p = Net::Ping->new();
592 $status = $p->ping($host); # || 0. Appears to set to undef rather than 0
593 print STDERR "**** $host is alive.\n" if $status; #print "$host is alive.\n" if $p->ping($host);
594 $p->close();
595 }
596 # return whether pinging was a success or failure
597 return $status;
598}
599
6001;
Note: See TracBrowser for help on using the repository browser.