| 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 |
|
|---|
| 28 | package servercontrol;
|
|---|
| 29 |
|
|---|
| 30 |
|
|---|
| 31 | use strict;
|
|---|
| 32 | no strict 'refs'; # allow filehandles to be variables and vice versa
|
|---|
| 33 | no strict 'subs'; # allow barewords (eg STDERR) as function arguments
|
|---|
| 34 |
|
|---|
| 35 | # Greenstone includes
|
|---|
| 36 | use printusage;
|
|---|
| 37 | use parse2;
|
|---|
| 38 |
|
|---|
| 39 |
|
|---|
| 40 | # The perl library imports below are used by deprecated methods config_old(), is_URL_active() and pingHost()
|
|---|
| 41 | # If the following library imports are not supported by your perl installation, comment out these
|
|---|
| 42 | # imports and move the methods config_old(), is_URL_active() and pingHost() out to a temporary file.
|
|---|
| 43 | use HTTP::Response;
|
|---|
| 44 | use LWP::Simple qw($ua !head); # import useragent object as $ua from the full LWP to use along with LWP::Simple
|
|---|
| 45 | # don't import LWP::Simple's head function by name since it can conflict with CGI:head())
|
|---|
| 46 | #use CGI qw(:standard); # then only CGI.pm defines a head()
|
|---|
| 47 | use Net::Ping;
|
|---|
| 48 | use URI;
|
|---|
| 49 |
|
|---|
| 50 |
|
|---|
| 51 | sub new
|
|---|
| 52 | {
|
|---|
| 53 | my $class = shift(@_);
|
|---|
| 54 |
|
|---|
| 55 | my ($qualified_collection, $site, $verbosity, $build_dir, $index_dir, $collect_dir, $library_url, $library_name) = @_;
|
|---|
| 56 |
|
|---|
| 57 | # library_url: to be specified on the cmdline if not using a GS-included web server
|
|---|
| 58 | # the GSDL_LIBRARY_URL env var is useful when running cmdline buildcol.pl in the linux package manager versions of GS3
|
|---|
| 59 |
|
|---|
| 60 | my $self = {'build_dir' => $build_dir,
|
|---|
| 61 | 'index_dir' => $index_dir,
|
|---|
| 62 | 'collect_dir' => $collect_dir,
|
|---|
| 63 | 'site' => $site,
|
|---|
| 64 | 'qualified_collection' => $qualified_collection,
|
|---|
| 65 | #'is_persistent_server' => undef,
|
|---|
| 66 | 'library_url' => $library_url || $ENV{'GSDL_LIBRARY_URL'} || undef, # to be specified on the cmdline if not using a GS-included web server
|
|---|
| 67 | 'library_name' => $library_name,
|
|---|
| 68 | #'gs_mode' => "gs2",
|
|---|
| 69 | 'verbosity' => $verbosity || 2
|
|---|
| 70 | };
|
|---|
| 71 |
|
|---|
| 72 | if ((defined $site) && ($site ne "")) { # GS3
|
|---|
| 73 | $self->{'gs_mode'} = "gs3";
|
|---|
| 74 | } else {
|
|---|
| 75 | $self->{'gs_mode'} = "gs2";
|
|---|
| 76 | }
|
|---|
| 77 |
|
|---|
| 78 | return bless($self, $class);
|
|---|
| 79 | }
|
|---|
| 80 |
|
|---|
| 81 | ## TODO: gsprintf to $self->{'out'} in these 2 print functions
|
|---|
| 82 | ## See buildcolutils.pm new() for setting up $out
|
|---|
| 83 |
|
|---|
| 84 | sub print_task_msg {
|
|---|
| 85 | my $self = shift(@_);
|
|---|
| 86 | my ($task_msg, $verbosity_setting) = @_;
|
|---|
| 87 |
|
|---|
| 88 | $verbosity_setting = $self->{'verbosity'} unless $verbosity_setting;
|
|---|
| 89 | #$verbosity_setting = 1 unless defined $verbosity;
|
|---|
| 90 | if($verbosity_setting >= 1) {
|
|---|
| 91 | print STDERR "\n";
|
|---|
| 92 | print STDERR "************************\n";
|
|---|
| 93 | print STDERR "* $task_msg\n";
|
|---|
| 94 | print STDERR "************************\n";
|
|---|
| 95 | }
|
|---|
| 96 | }
|
|---|
| 97 |
|
|---|
| 98 | # Prints messages if the verbosity is right. Does not add new lines.
|
|---|
| 99 | sub print_msg {
|
|---|
| 100 | my $self = shift(@_);
|
|---|
| 101 | my ($msg, $min_verbosity, $verbosity_setting) = @_;
|
|---|
| 102 |
|
|---|
| 103 | # only display error messages if the current
|
|---|
| 104 | # verbosity setting >= the minimum verbosity level
|
|---|
| 105 | # needed for that message to be displayed.
|
|---|
| 106 |
|
|---|
| 107 | $verbosity_setting = $self->{'verbosity'} unless defined $verbosity_setting;
|
|---|
| 108 | $min_verbosity = 1 unless defined $min_verbosity;
|
|---|
| 109 | if($verbosity_setting >= $min_verbosity) { # by default display all 1 messages
|
|---|
| 110 | print STDERR "$msg";
|
|---|
| 111 | }
|
|---|
| 112 | }
|
|---|
| 113 |
|
|---|
| 114 | # Method to send a command to a GS2 or GS3 library_URL
|
|---|
| 115 | # the commands used in this script can be activate, deactivate, ping,
|
|---|
| 116 | # and is-persistent (is-persistent only implemented for GS2).
|
|---|
| 117 | sub config {
|
|---|
| 118 | my $self = shift(@_);
|
|---|
| 119 | my ($command, $check_message_against_regex, $expected_error_code, $silent) = @_;
|
|---|
| 120 |
|
|---|
| 121 | my $library_url = $self->get_library_URL(); #$self->{'library_url'};
|
|---|
| 122 |
|
|---|
| 123 |
|
|---|
| 124 | # Gatherer.java's configGS3Server doesn't use the site variable
|
|---|
| 125 | # so we don't have to either
|
|---|
| 126 |
|
|---|
| 127 | # for GS2, getting the HTTP status isn't enough, we need to read the output
|
|---|
| 128 | # since this is what CollectionManager.config() stipulates.
|
|---|
| 129 | # Using LWP::UserAgent::get($url) for this
|
|---|
| 130 |
|
|---|
| 131 | if(!defined $library_url) {
|
|---|
| 132 | return 0;
|
|---|
| 133 | }
|
|---|
| 134 | else {
|
|---|
| 135 | # ampersands need to be escaped
|
|---|
| 136 | # - with single quotes around it for linux for the cmd to run in bash subshell
|
|---|
| 137 | # - with a ^ before it on windows for the cmd to run in a DOS prompt subshell
|
|---|
| 138 | # - or the entire wget command should be nested in double quotes (single quotes don't work on windows)
|
|---|
| 139 | my $wgetCommand = $command;
|
|---|
| 140 |
|
|---|
| 141 | my $wget_file_path = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "wget");
|
|---|
| 142 | my $tmpfilename = &util::get_tmp_filename(".html"); # random file name with html extension in tmp location in which we'll store the HTML page retrieved by wget
|
|---|
| 143 |
|
|---|
| 144 | # https://www.gnu.org/software/wget/manual/wget.html
|
|---|
| 145 | # output-document set to - (STDOUT), so page is streamed to STDOUT
|
|---|
| 146 | # timeout: 5 seconds, tries: 1
|
|---|
| 147 | # wget sends status information and response code to STDERR, so redirect it to STDOUT
|
|---|
| 148 | # Searching for "perl backtick operator redirect stderr to stdout":
|
|---|
| 149 | # http://www.perlmonks.org/?node=How%20can%20I%20capture%20STDERR%20from%20an%20external%20command%3F
|
|---|
| 150 | ##$wgetCommand = "\"$wget_file_path\" --spider -T 5 -t 1 \"$library_url$wgetCommand\" 2>&1"; # won't save page
|
|---|
| 151 | #$wgetCommand = "\"$wget_file_path\" --output-document=- -T 5 -t 1 \"$library_url$wgetCommand\" 2>&1"; # THIS CAN MIX UP STDERR WITH STDOUT IN THE VERY LINE WE REGEX TEST AGAINST EXPECTED OUTPUT!!
|
|---|
| 152 | $wgetCommand = "\"$wget_file_path\" --output-document=$tmpfilename -T 5 -t 1 \"$library_url$wgetCommand\" 2>&1"; # keep stderr (response code, response_content) separate from html page content
|
|---|
| 153 |
|
|---|
| 154 | ##print STDERR "@@@@ $wgetCommand\n";
|
|---|
| 155 |
|
|---|
| 156 | my $response_content;
|
|---|
| 157 | my $response_code = undef;
|
|---|
| 158 | #my $response_content = `$wgetCommand`; # Dr Bainbridge advises against using backticks for running a process. If capturing std output, use open():
|
|---|
| 159 | if (open(PIN, "$wgetCommand |")) {
|
|---|
| 160 | while (defined (my $perl_output_line = <PIN>)) {
|
|---|
| 161 | $response_content = $response_content . $perl_output_line;
|
|---|
| 162 | }
|
|---|
| 163 | close(PIN);
|
|---|
| 164 | } else {
|
|---|
| 165 | print STDERR "servercontrol.pm::config() failed to run $wgetCommand\n";
|
|---|
| 166 | }
|
|---|
| 167 |
|
|---|
| 168 |
|
|---|
| 169 | my @lines = split( /\n/, $response_content );
|
|---|
| 170 | foreach my $line (@lines) {
|
|---|
| 171 | #print STDERR "@@@@ LINE: $line\n";
|
|---|
| 172 | if($line =~ m@failed: Connection timed out.$@) { # linux
|
|---|
| 173 | $response_code = "failed: Connection timed out.";
|
|---|
| 174 | last; # break keyword in perl = last
|
|---|
| 175 | }
|
|---|
| 176 | elsif($line =~ m@Giving up.$@) { # windows (unless -T 5 -t 1 is not passed in)
|
|---|
| 177 | $response_code = "failed: Giving up.";
|
|---|
| 178 | last; # break keyword in perl = last
|
|---|
| 179 | }
|
|---|
| 180 | elsif($line =~ m@failed: Connection refused.$@) {
|
|---|
| 181 | $response_code = "failed: Connection refused.";
|
|---|
| 182 | last; # break keyword in perl = last
|
|---|
| 183 | }
|
|---|
| 184 | elsif($line =~ m@HTTP request sent, @) {
|
|---|
| 185 | $response_code = $line;
|
|---|
| 186 | $response_code =~ s@[^\d]*(.*)$@$1@;
|
|---|
| 187 | last;
|
|---|
| 188 | }
|
|---|
| 189 | }
|
|---|
| 190 |
|
|---|
| 191 | if($command =~ m@ping@ && $response_code =~ m@failed: (Connection refused|Giving up)@) {
|
|---|
| 192 | # server not running
|
|---|
| 193 | $self->print_msg("*** Server not running. $library_url$command\n", 3);
|
|---|
| 194 | &FileUtils::removeFiles($tmpfilename); # get rid of the ping response's temporary html file we downloaded
|
|---|
| 195 | return 0;
|
|---|
| 196 | }
|
|---|
| 197 | if($response_code && $response_code eq "200 OK") {
|
|---|
| 198 | $self->print_msg("*** Command $library_url$command\n", 3);
|
|---|
| 199 | $self->print_msg("*** HTTP Response Status: $response_code - Complete.", 3);
|
|---|
| 200 |
|
|---|
| 201 | # check the page content is as expected
|
|---|
| 202 | #my $resultstr = $response_content;
|
|---|
| 203 |
|
|---|
| 204 | open(FIN,"<$tmpfilename") or die "servercontrol.pm: Unable to open $tmpfilename to read ping response page...ERROR: $!\n";
|
|---|
| 205 | my $resultstr;
|
|---|
| 206 | # Read in the entire contents of the file in one hit
|
|---|
| 207 | sysread(FIN, $resultstr, -s FIN);
|
|---|
| 208 | close(FIN);
|
|---|
| 209 | &FileUtils::removeFiles($tmpfilename); # get rid of the ping response's temporary html file we downloaded
|
|---|
| 210 |
|
|---|
| 211 |
|
|---|
| 212 | #$resultstr =~ s@.*gs_content\"\>@@s; ## only true for default library servlet
|
|---|
| 213 | #$resultstr =~ s@</div>.*@@s;
|
|---|
| 214 | if($resultstr =~ m/$check_message_against_regex/) {
|
|---|
| 215 | $self->print_msg(" Response as expected.\n", 3);
|
|---|
| 216 | $self->print_msg("@@@@@@ Got result:\n$resultstr\n", 4);
|
|---|
| 217 | return 1;
|
|---|
| 218 | } else {
|
|---|
| 219 | # if we expect the collection to be inactive, then we'd be in silent mode: if so,
|
|---|
| 220 | # don't print out the "ping did not succeed" response, but print out any other messages
|
|---|
| 221 |
|
|---|
| 222 | # So we only suppress the ping col "did not succeed" response if we're in silent mode
|
|---|
| 223 | # But if any message other than ping "did not succeed" is returned, we always print it
|
|---|
| 224 | if($resultstr !~ m/did not succeed/ || !$silent) {
|
|---|
| 225 | $self->print_msg("\n\tBUT: command $library_url$command response UNEXPECTED.\n", 3);
|
|---|
| 226 | $self->print_msg("*** Got message:\n$response_content.\n", 4);
|
|---|
| 227 | $self->print_msg("*** Got result:\n$resultstr\n", 3);
|
|---|
| 228 | }
|
|---|
| 229 | return 0; # ping on a collection may "not succeed."
|
|---|
| 230 | }
|
|---|
| 231 | }
|
|---|
| 232 | elsif($response_code && $response_code =~ m@^(4|5)\d\d@) { # client side errors start with 4xx, server side with 5xx
|
|---|
| 233 | # check the page content is as expected
|
|---|
| 234 | if(defined $expected_error_code && $response_code =~ m@^$expected_error_code@) {
|
|---|
| 235 | $self->print_msg(" Response status $response_code as expected.\n", 3);
|
|---|
| 236 | } else {
|
|---|
| 237 | $self->print_msg("*** Command $library_url$command\n");
|
|---|
| 238 | $self->print_msg("*** Unexpected error type 1. HTTP Response Status: $response_code - Failed.\n");
|
|---|
| 239 | }
|
|---|
| 240 | return 0; # return false, since the response_code was an error, expected or not
|
|---|
| 241 | }
|
|---|
| 242 | else { # also if response_code is still undefined, as can happen with connection timing out
|
|---|
| 243 | $self->print_msg("*** Command $library_url$command\n");
|
|---|
| 244 | if(defined $response_code) {
|
|---|
| 245 | $self->print_msg("*** Unexpected error type 2. HTTP Response Status: $response_code - Failed.\n");
|
|---|
| 246 | } else {
|
|---|
| 247 | $self->print_msg("*** Unexpected error type 3. Failed:\n\n$response_content\n\n");
|
|---|
| 248 | }
|
|---|
| 249 | return 0;
|
|---|
| 250 | }
|
|---|
| 251 | #print STDERR "********** WgetCommand: $wgetCommand\n\n";
|
|---|
| 252 | #print STDERR "********** Response_content:\n$response_content\n\n";
|
|---|
| 253 | #print STDERR "********** Response_CODE: $response_code\n";
|
|---|
| 254 |
|
|---|
| 255 | }
|
|---|
| 256 | }
|
|---|
| 257 |
|
|---|
| 258 | sub deactivate_collection {
|
|---|
| 259 | my $self = shift(@_);
|
|---|
| 260 |
|
|---|
| 261 | my $gs_mode = $self->{'gs_mode'};
|
|---|
| 262 | my $qualified_collection = $self->{'qualified_collection'};
|
|---|
| 263 |
|
|---|
| 264 | if($gs_mode eq "gs2") {
|
|---|
| 265 | my $DEACTIVATE_COMMAND = "?a=config&cmd=release-collection&c=";
|
|---|
| 266 | my $check_message_against_regex = q/configured release-collection/;
|
|---|
| 267 | $self->config($DEACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
|
|---|
| 268 | }
|
|---|
| 269 | elsif ($gs_mode eq "gs3") {
|
|---|
| 270 | my $DEACTIVATE_COMMAND = "?a=s&sa=d&st=collection&sn=";
|
|---|
| 271 | my $check_message_against_regex = "collection: $qualified_collection deactivated";
|
|---|
| 272 | $self->config($DEACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
|
|---|
| 273 | }
|
|---|
| 274 | }
|
|---|
| 275 |
|
|---|
| 276 | sub activate_collection {
|
|---|
| 277 | my $self = shift(@_);
|
|---|
| 278 |
|
|---|
| 279 | my $gs_mode = $self->{'gs_mode'};
|
|---|
| 280 | my $qualified_collection = $self->{'qualified_collection'};
|
|---|
| 281 |
|
|---|
| 282 | if($gs_mode eq "gs2") {
|
|---|
| 283 | my $ACTIVATE_COMMAND = "?a=config&cmd=add-collection&c=";
|
|---|
| 284 | my $check_message_against_regex = q/configured add-collection/;
|
|---|
| 285 | $self->config($ACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
|
|---|
| 286 | }
|
|---|
| 287 | elsif ($gs_mode eq "gs3") {
|
|---|
| 288 | my $ACTIVATE_COMMAND = "?a=s&sa=a&st=collection&sn=";
|
|---|
| 289 | my $check_message_against_regex = "collection: $qualified_collection activated";
|
|---|
| 290 | $self->config($ACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
|
|---|
| 291 | }
|
|---|
| 292 | }
|
|---|
| 293 |
|
|---|
| 294 | sub ping {
|
|---|
| 295 | my $self = shift(@_);
|
|---|
| 296 | my $command = shift(@_);
|
|---|
| 297 | my $silent = shift(@_);
|
|---|
| 298 |
|
|---|
| 299 | # If the GS server is not running, we *expect* to see a "500" status code.
|
|---|
| 300 | # If the GS server is running, then "Ping" ... "succeeded" is expected on success.
|
|---|
| 301 | # When pinging an inactive collection, it will say it did "not succeed". This is
|
|---|
| 302 | # a message of interest to return.
|
|---|
| 303 | my $check_responsemsg_against_regex = q/(succeeded)/;
|
|---|
| 304 | my $expected_error_code = 500;
|
|---|
| 305 |
|
|---|
| 306 | $self->print_msg("*** COMMAND WAS: |$command| ***\n", 4);
|
|---|
| 307 |
|
|---|
| 308 | return $self->config($command, $check_responsemsg_against_regex, $expected_error_code, $silent);
|
|---|
| 309 | }
|
|---|
| 310 |
|
|---|
| 311 | # send a pingaction to the GS library. General server-level ping.
|
|---|
| 312 | sub ping_library {
|
|---|
| 313 | my $self = shift(@_);
|
|---|
| 314 |
|
|---|
| 315 | my $gs_mode = $self->{'gs_mode'};
|
|---|
| 316 |
|
|---|
| 317 | my $command = "";
|
|---|
| 318 | if($gs_mode eq "gs2") {
|
|---|
| 319 | $command = "?a=ping";
|
|---|
| 320 | }
|
|---|
| 321 | elsif ($gs_mode eq "gs3") {
|
|---|
| 322 | $command = "?a=s&sa=ping";
|
|---|
| 323 | }
|
|---|
| 324 | return $self->ping($command);
|
|---|
| 325 | }
|
|---|
| 326 |
|
|---|
| 327 |
|
|---|
| 328 | # send a pingaction to a collection in GS library to check if it's active
|
|---|
| 329 | sub ping_library_collection {
|
|---|
| 330 | my $self = shift(@_);
|
|---|
| 331 | my $silent = shift(@_);
|
|---|
| 332 |
|
|---|
| 333 | my $gs_mode = $self->{'gs_mode'};
|
|---|
| 334 | my $qualified_collection = $self->{'qualified_collection'};
|
|---|
| 335 |
|
|---|
| 336 | my $command = "";
|
|---|
| 337 | if($gs_mode eq "gs2") {
|
|---|
| 338 | $command = "?a=ping&c=$qualified_collection";
|
|---|
| 339 | }
|
|---|
| 340 | elsif ($gs_mode eq "gs3") {
|
|---|
| 341 | $command = "?a=s&sa=ping&st=collection&sn=$qualified_collection";
|
|---|
| 342 | }
|
|---|
| 343 | return $self->ping($command, $silent);
|
|---|
| 344 | }
|
|---|
| 345 |
|
|---|
| 346 | # return true if server is persistent, by calling is-persistent on library_url
|
|---|
| 347 | # this is only for GS2, since the GS3 server is always persistent
|
|---|
| 348 | sub is_persistent {
|
|---|
| 349 | my $self = shift(@_);
|
|---|
| 350 |
|
|---|
| 351 | if($self->{'gs_mode'} eq "gs3") { # GS3 server is always persistent
|
|---|
| 352 | return 1;
|
|---|
| 353 | }
|
|---|
| 354 |
|
|---|
| 355 | my $command = "?a=is-persistent";
|
|---|
| 356 | my $check_responsemsg_against_regex = q/true/; # isPersistent: true versus isPersistent: false
|
|---|
| 357 | return $self->config($command, $check_responsemsg_against_regex);
|
|---|
| 358 | }
|
|---|
| 359 |
|
|---|
| 360 | sub set_library_URL {
|
|---|
| 361 | my $self = shift(@_);
|
|---|
| 362 | my $library_url = shift(@_);
|
|---|
| 363 | $self->{'library_url'} = $library_url;
|
|---|
| 364 | }
|
|---|
| 365 |
|
|---|
| 366 | sub get_library_URL {
|
|---|
| 367 | my $self = shift(@_);
|
|---|
| 368 |
|
|---|
| 369 | # For web servers that are external to a Greenstone installation,
|
|---|
| 370 | # the user can pass in their web server's library URL.
|
|---|
| 371 | if($self->{'library_url'}) {
|
|---|
| 372 | return $self->{'library_url'};
|
|---|
| 373 | }
|
|---|
| 374 |
|
|---|
| 375 | # For web servers included with GS (like tomcat for GS3 and server.exe
|
|---|
| 376 | # and apache for GS2), we work out the library URL:
|
|---|
| 377 | my ($gs_mode, $lib_name); # gs_mode can be gs3 or gs2, lib_name is the custom servlet name
|
|---|
| 378 | $gs_mode = $self->{'gs_mode'};
|
|---|
| 379 | $lib_name = $self->{'library_name'};
|
|---|
| 380 |
|
|---|
| 381 | # If we get here, we are dealing with a server included with GS.
|
|---|
| 382 | # For GS3, we ask ant for the library URL.
|
|---|
| 383 | # For GS2, we derive the URL from the llssite.cfg file.
|
|---|
| 384 |
|
|---|
| 385 | my $url = undef;
|
|---|
| 386 |
|
|---|
| 387 | if($gs_mode eq "gs2") {
|
|---|
| 388 | my $llssite_cfg = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "llssite.cfg");
|
|---|
| 389 |
|
|---|
| 390 | if(-f $llssite_cfg) {
|
|---|
| 391 | # check llssite.cfg for line with url property
|
|---|
| 392 | # for server.exe also need to use portnumber and enterlib properties
|
|---|
| 393 |
|
|---|
| 394 | # Read in the entire contents of the file in one hit
|
|---|
| 395 | if (!open (FIN, $llssite_cfg)) {
|
|---|
| 396 | $self->print_msg("activate.pl::get_library_URL failed to open $llssite_cfg ($!)\n");
|
|---|
| 397 | return undef;
|
|---|
| 398 | }
|
|---|
| 399 |
|
|---|
| 400 | my $contents;
|
|---|
| 401 | sysread(FIN, $contents, -s FIN);
|
|---|
| 402 | close(FIN);
|
|---|
| 403 |
|
|---|
| 404 | my @lines = split(/[\n\r]+/, $contents); # split on carriage-returns and/or linefeeds
|
|---|
| 405 | my $enterlib = "";
|
|---|
| 406 | my $portnumber = "8282"; # will remain empty (implicit port 80) unless it's specifically been assigned
|
|---|
| 407 |
|
|---|
| 408 | foreach my $line (@lines) {
|
|---|
| 409 | if($line =~ m/^url=(.*)$/) {
|
|---|
| 410 | $url = $1;
|
|---|
| 411 | } elsif($line =~ m/^enterlib=(.*)$/) {
|
|---|
| 412 | $enterlib = $1;
|
|---|
| 413 | } elsif($line =~ m/^portnumber=(.*)$/) {
|
|---|
| 414 | $portnumber = $1;
|
|---|
| 415 | }
|
|---|
| 416 | }
|
|---|
| 417 |
|
|---|
| 418 | if(!$url) {
|
|---|
| 419 | return undef;
|
|---|
| 420 | }
|
|---|
| 421 | elsif($url eq "URL_pending") { # library is not running
|
|---|
| 422 | # do not process url=URL_pending in the file, since for server.exe
|
|---|
| 423 | # this just means the Enter Library button hasn't been pressed yet
|
|---|
| 424 | $url = undef;
|
|---|
| 425 | }
|
|---|
| 426 | else {
|
|---|
| 427 | # In the case of server.exe, need to do extra work to get the proper URL
|
|---|
| 428 | # But first, need to know whether we're indeed dealing with server.exe:
|
|---|
| 429 |
|
|---|
| 430 | # compare the URL's domain to the full URL
|
|---|
| 431 | # E.g. for http://localhost:8383/greenstone3/cgi-bin, the domain is localhost:8383
|
|---|
| 432 | my $uri = URI->new( $url );
|
|---|
| 433 | my $host = $uri->host;
|
|---|
| 434 | #print STDERR "@@@@@ host: $host\n";
|
|---|
| 435 | if($url =~ m/https?:\/\/$host(\/)?$/) {
|
|---|
| 436 | #if($url !~ m/https?:\/\/$host:$portnumber(\/)?/ || $url =~ m/https?:\/\/$host(\/)?$/) {
|
|---|
| 437 | # (if the URL does not contain the portnumber, OR if the port is implicitly 80 and)
|
|---|
| 438 | # If the domain with http:// prefix is completely the same as the URL, assume server.exe
|
|---|
| 439 | # then the actual URL is the result of suffixing the port and enterlib properties in llssite.cfg
|
|---|
| 440 | $url = $url.":".$portnumber.$enterlib;
|
|---|
| 441 | } # else, apache web server
|
|---|
| 442 |
|
|---|
| 443 | }
|
|---|
| 444 | }
|
|---|
| 445 | } elsif($gs_mode eq "gs3") {
|
|---|
| 446 | # Either check build.properties for tomcat.server, tomcat.port and app.name (and default servlet name).
|
|---|
| 447 | # app.name is stored in app.path by build.xml. Need to move app.name in build.properties from build.xml
|
|---|
| 448 |
|
|---|
| 449 | # Or, run the new target get-default-servlet-url
|
|---|
| 450 | # the output can look like:
|
|---|
| 451 | #
|
|---|
| 452 | # Buildfile: build.xml
|
|---|
| 453 | # [echo] os.name: Windows Vista
|
|---|
| 454 | #
|
|---|
| 455 | # get-default-servlet-url:
|
|---|
| 456 | # [echo] http://localhost:8383/greenstone3/library
|
|---|
| 457 | # BUILD SUCCESSFUL
|
|---|
| 458 | # Total time: 0 seconds
|
|---|
| 459 |
|
|---|
| 460 | #my $output = qx/ant get-default-servlet-url/; # backtick operator, to get STDOUT (else 2>&1)
|
|---|
| 461 | # - see http://stackoverflow.com/questions/799968/whats-the-difference-between-perls-backticks-system-and-exec
|
|---|
| 462 |
|
|---|
| 463 | # The get-default-servlet-url ant target can be run from anywhere by specifying the
|
|---|
| 464 | # location of GS3's ant build.xml buildfile. Activate.pl can be run from anywhere for GS3
|
|---|
| 465 | # GSDL3SRCHOME will be set for GS3 by gs3-setup.sh, a step that would have been necessary
|
|---|
| 466 | # to run the activate.pl script in the first place
|
|---|
| 467 |
|
|---|
| 468 | my $full_build_xml = &FileUtils::javaFilenameConcatenate($ENV{'GSDL3SRCHOME'},"build.xml");
|
|---|
| 469 | my $perl_command = "ant -buildfile \"$full_build_xml\" get-default-servlet-url";
|
|---|
| 470 |
|
|---|
| 471 | if (open(PIN, "$perl_command |")) {
|
|---|
| 472 | while (defined (my $perl_output_line = <PIN>)) {
|
|---|
| 473 | if($perl_output_line =~ m@(https?):\/\/(\S*)@) { # grab all the non-whitespace chars
|
|---|
| 474 | $url="$1://".$2; # preserve the http protocol #$url="http://".$1;
|
|---|
| 475 | }
|
|---|
| 476 | }
|
|---|
| 477 | close(PIN);
|
|---|
| 478 |
|
|---|
| 479 | if(defined $lib_name) { # url won't be undef
|
|---|
| 480 | # replace the servlet_name portion of the url found, with the given library_name
|
|---|
| 481 | $url =~ s@/[^/]*$@/$lib_name@;
|
|---|
| 482 | }
|
|---|
| 483 | } else {
|
|---|
| 484 | $self->print_msg("activate.pl::get_library_URL: Failed to run $perl_command to work out library URL for $gs_mode\n");
|
|---|
| 485 | }
|
|---|
| 486 | }
|
|---|
| 487 |
|
|---|
| 488 | # either the url is still undef or it is now set
|
|---|
| 489 | #print STDERR "\n@@@@@ final URL:|$url|\n" if $url;
|
|---|
| 490 | #print STDERR "\n@@@@@ URL still undef\n" if !$url;
|
|---|
| 491 |
|
|---|
| 492 | $self->{'library_url'} = $url;
|
|---|
| 493 | return $url;
|
|---|
| 494 | }
|
|---|
| 495 |
|
|---|
| 496 |
|
|---|
| 497 | sub do_deactivate {
|
|---|
| 498 | my $self = shift(@_);
|
|---|
| 499 |
|
|---|
| 500 | # 1. Get library URL
|
|---|
| 501 |
|
|---|
| 502 | # For web servers that are external to a Greenstone installation,
|
|---|
| 503 | # the user can pass in their web server's library URL.
|
|---|
| 504 | # For web servers included with GS (like tomcat for GS3 and server.exe
|
|---|
| 505 | # and apache for GS2), we work out the library URL:
|
|---|
| 506 |
|
|---|
| 507 | # Can't do $self->{'library_url'}, because it may not yet be set
|
|---|
| 508 | my $library_url = $self->get_library_URL(); # returns undef if no valid server URL
|
|---|
| 509 |
|
|---|
| 510 | if(!defined $library_url) { # undef if no valid server URL
|
|---|
| 511 | return; # can't do any deactivation without a valid server URL
|
|---|
| 512 | }
|
|---|
| 513 |
|
|---|
| 514 | my $is_persistent_server = $self->{'is_persistent_server'};
|
|---|
| 515 | my $qualified_collection = $self->{'qualified_collection'};
|
|---|
| 516 |
|
|---|
| 517 | # CollectionManager's installCollection phase in GLI
|
|---|
| 518 | # 2. Ping the library URL, and if it's a persistent server and running, release the collection
|
|---|
| 519 |
|
|---|
| 520 | $self->print_msg("Pinging $library_url\n");
|
|---|
| 521 | if ($self->ping_library()) { # server running
|
|---|
| 522 |
|
|---|
| 523 | # server is running, so release the collection if
|
|---|
| 524 | # the server is persistent and the collection is active
|
|---|
| 525 |
|
|---|
| 526 | # don't need to work out persistency of server more than once, since the libraryURL hasn't changed
|
|---|
| 527 | if (!defined $is_persistent_server) {
|
|---|
| 528 | $self->print_msg("Checking if Greenstone server is persistent\n");
|
|---|
| 529 | $is_persistent_server = $self->is_persistent();
|
|---|
| 530 | $self->{'is_persistent_server'} = $is_persistent_server;
|
|---|
| 531 | }
|
|---|
| 532 |
|
|---|
| 533 | if ($is_persistent_server) { # only makes sense to issue activate and deactivate cmds to a persistent server
|
|---|
| 534 |
|
|---|
| 535 | $self->print_msg("Checking if the collection $qualified_collection is already active\n");
|
|---|
| 536 | my $collection_active = $self->ping_library_collection();
|
|---|
| 537 |
|
|---|
| 538 | if ($collection_active) {
|
|---|
| 539 | $self->print_msg("De-activating collection $qualified_collection\n");
|
|---|
| 540 | $self->deactivate_collection();
|
|---|
| 541 | }
|
|---|
| 542 | else {
|
|---|
| 543 | $self->print_msg("Collection is not active => No need to deactivate\n");
|
|---|
| 544 | }
|
|---|
| 545 | }
|
|---|
| 546 | else {
|
|---|
| 547 | $self->print_msg("Server is not persistent => No need to deactivate collection\n");
|
|---|
| 548 | }
|
|---|
| 549 | }
|
|---|
| 550 | else {
|
|---|
| 551 | $self->print_msg("No response to Ping => Taken to mean server is not running\n");
|
|---|
| 552 | }
|
|---|
| 553 |
|
|---|
| 554 | return $is_persistent_server;
|
|---|
| 555 | }
|
|---|
| 556 |
|
|---|
| 557 | sub do_activate {
|
|---|
| 558 | my $self = shift @_;
|
|---|
| 559 |
|
|---|
| 560 | my $library_url = $self->get_library_URL(); # Can't do $self->{'library_url'}; as it may not be set yet
|
|---|
| 561 |
|
|---|
| 562 | if(!defined $library_url) { # undef if no valid server URL
|
|---|
| 563 | return; # nothing to activate if without valid server URL
|
|---|
| 564 | }
|
|---|
| 565 |
|
|---|
| 566 | my $is_persistent_server = $self->{'is_persistent_server'};
|
|---|
| 567 | my $qualified_collection = $self->{'qualified_collection'};
|
|---|
| 568 |
|
|---|
| 569 | $self->print_msg("Pinging $library_url\n");
|
|---|
| 570 | if ($self->ping_library()) { # server running
|
|---|
| 571 |
|
|---|
| 572 | # don't need to work out persistency of server more than once, since the libraryURL hasn't changed
|
|---|
| 573 | if (!defined $is_persistent_server) {
|
|---|
| 574 | $self->print_msg("Checking if Greenstone server is persistent\n");
|
|---|
| 575 | $is_persistent_server = $self->is_persistent();
|
|---|
| 576 | $self->{'is_persistent_server'} = $is_persistent_server;
|
|---|
| 577 | }
|
|---|
| 578 |
|
|---|
| 579 | if ($is_persistent_server) { # persistent server, so can try activating collection
|
|---|
| 580 |
|
|---|
| 581 | $self->print_msg("Checking if the collection $qualified_collection is not already active\n");
|
|---|
| 582 |
|
|---|
| 583 | # Since we could have deactivated the collection at this point,
|
|---|
| 584 | # it is likely that it is not yet active. When pinging the collection
|
|---|
| 585 | # a "ping did not succeed" message is expected, therefore tell the ping
|
|---|
| 586 | # to proceed silently
|
|---|
| 587 | my $silent = 1;
|
|---|
| 588 | my $collection_active = $self->ping_library_collection($silent);
|
|---|
| 589 |
|
|---|
| 590 | if (!$collection_active) {
|
|---|
| 591 | $self->print_msg(" Collection is not active.\n");
|
|---|
| 592 | $self->print_msg("Activating collection $qualified_collection\n");
|
|---|
| 593 | $self->activate_collection();
|
|---|
| 594 |
|
|---|
| 595 | # unless an error occurred, the collection should now be active:
|
|---|
| 596 | $collection_active = $self->ping_library_collection(); # not silent if ping did not succeed
|
|---|
| 597 | if(!$collection_active) {
|
|---|
| 598 | $self->print_msg("ERROR: collection $qualified_collection did not get activated\n");
|
|---|
| 599 | }
|
|---|
| 600 | }
|
|---|
| 601 | else {
|
|---|
| 602 | $self->print_msg("Collection is already active => No need to activate\n");
|
|---|
| 603 | }
|
|---|
| 604 | }
|
|---|
| 605 | else {
|
|---|
| 606 | $self->print_msg("Server is not persistent => No need to activate collection\n");
|
|---|
| 607 | }
|
|---|
| 608 | }
|
|---|
| 609 | else {
|
|---|
| 610 | $self->print_msg("No response to Ping => Taken to mean server is not running\n");
|
|---|
| 611 | }
|
|---|
| 612 |
|
|---|
| 613 | return $is_persistent_server;
|
|---|
| 614 | }
|
|---|
| 615 |
|
|---|
| 616 |
|
|---|
| 617 | #########################################################
|
|---|
| 618 | ### UNUSED METHODS - CAN BE HANDY
|
|---|
| 619 |
|
|---|
| 620 |
|
|---|
| 621 | # This method uses the perl libraries we're advised to use in place of wget for pinging and retrieving web
|
|---|
| 622 | # pages. The problem is that not all perl installations may support these libraries. So we now use the new
|
|---|
| 623 | # config() method further above, which uses the wget included in Greenstone binary installations.
|
|---|
| 624 | # If the library imports at page top conflict, comment out those imports and move the methods config_old(),
|
|---|
| 625 | # is_URL_active() and pingHost() out to a temporary file.
|
|---|
| 626 | #
|
|---|
| 627 | # If for some reason you can't use wget, then rename the config() method to config_old(), and rename the
|
|---|
| 628 | # method below to config() and things should work as before.
|
|---|
| 629 | sub config_old {
|
|---|
| 630 | my $self = shift(@_);
|
|---|
| 631 | my ($command, $check_message_against_regex, $expected_error_code, $silent) = @_;
|
|---|
| 632 |
|
|---|
| 633 | my $library_url = $self->get_library_URL(); #$self->{'library_url'};
|
|---|
| 634 |
|
|---|
| 635 |
|
|---|
| 636 | # Gatherer.java's configGS3Server doesn't use the site variable
|
|---|
| 637 | # so we don't have to either
|
|---|
| 638 |
|
|---|
| 639 | # for GS2, getting the HTTP status isn't enough, we need to read the output
|
|---|
| 640 | # since this is what CollectionManager.config() stipulates.
|
|---|
| 641 | # Using LWP::UserAgent::get($url) for this
|
|---|
| 642 |
|
|---|
| 643 | if(!defined $library_url) {
|
|---|
| 644 | return 0;
|
|---|
| 645 | }
|
|---|
| 646 | else {
|
|---|
| 647 | $ua->timeout(5); # set LWP useragent to 5s max timeout for testing the URL
|
|---|
| 648 | # Need to set this, else it takes I don't know how long to timeout
|
|---|
| 649 | # http://www.perlmonks.org/?node_id=618534
|
|---|
| 650 |
|
|---|
| 651 | # http://search.cpan.org/~gaas/libwww-perl-6.04/lib/LWP/UserAgent.pm
|
|---|
| 652 | # use LWP::UserAgent's get($url) since it returns an HTTP::Response code
|
|---|
| 653 |
|
|---|
| 654 | my $response_obj = $ua->get($library_url.$command);
|
|---|
| 655 |
|
|---|
| 656 | # $response_obj->content stores the content and $response_obj->code the HTTP response code
|
|---|
| 657 | my $response_code = $response_obj->code();
|
|---|
| 658 |
|
|---|
| 659 | if(LWP::Simple::is_success($response_code)) {# $response_code eq RC_OK) { # LWP::Simple::is_success($response_code)
|
|---|
| 660 | $self->print_msg("*** Command $library_url$command\n", 3);
|
|---|
| 661 | $self->print_msg("*** HTTP Response Status: $response_code - Complete.", 3);
|
|---|
| 662 |
|
|---|
| 663 | # check the page content is as expected
|
|---|
| 664 | my $response_content = $response_obj->content;
|
|---|
| 665 | my $resultstr = $response_content;
|
|---|
| 666 | $resultstr =~ s@.*gs_content\"\>@@s;
|
|---|
| 667 | $resultstr =~ s@</div>.*@@s;
|
|---|
| 668 |
|
|---|
| 669 | if($resultstr =~ m/$check_message_against_regex/) {#if($response_content =~ m/$check_message_against_regex/) {
|
|---|
| 670 | $self->print_msg(" Response as expected.\n", 3);
|
|---|
| 671 | $self->print_msg("@@@@@@ Got result:\n$resultstr\n", 4);
|
|---|
| 672 | return 1;
|
|---|
| 673 | } else {
|
|---|
| 674 | # if we expect the collection to be inactive, then we'd be in silent mode: if so,
|
|---|
| 675 | # don't print out the "ping did not succeed" response, but print out any other messages
|
|---|
| 676 |
|
|---|
| 677 | # So we only suppress the ping col "did not succeed" response if we're in silent mode
|
|---|
| 678 | # But if any message other than ping "did not succeed" is returned, we always print it
|
|---|
| 679 | if($resultstr !~ m/did not succeed/ || !$silent) {#if($response_content !~ m/did not succeed/ || !$silent) {
|
|---|
| 680 | $self->print_msg("\n\tBUT: command $library_url$command response UNEXPECTED.\n", 3);
|
|---|
| 681 | $self->print_msg("*** Got message:\n$response_content.\n", 4);
|
|---|
| 682 | $self->print_msg("*** Got result:\n$resultstr\n", 3);
|
|---|
| 683 | }
|
|---|
| 684 | return 0; # ping on a collection may "not succeed."
|
|---|
| 685 | }
|
|---|
| 686 | }
|
|---|
| 687 | elsif(LWP::Simple::is_error($response_code)) { # method exported by LWP::Simple, along with HTTP::Status constants
|
|---|
| 688 | # check the page content is as expected
|
|---|
| 689 | if(defined $expected_error_code && $response_code == $expected_error_code) {
|
|---|
| 690 | $self->print_msg(" Response status $response_code as expected.\n", 3);
|
|---|
| 691 | } else {
|
|---|
| 692 | $self->print_msg("*** Command $library_url$command\n");
|
|---|
| 693 | $self->print_msg("*** Unexpected error. HTTP Response Status: $response_code - Failed.\n");
|
|---|
| 694 | }
|
|---|
| 695 | return 0; # return false, since the response_code was an error, expected or not
|
|---|
| 696 | }
|
|---|
| 697 | else {
|
|---|
| 698 | $self->print_msg("*** Command $library_url$command\n");
|
|---|
| 699 | $self->print_msg("*** Unexpected error. HTTP Response Status: $response_code - Failed.\n");
|
|---|
| 700 | return 0;
|
|---|
| 701 | }
|
|---|
| 702 | }
|
|---|
| 703 | }
|
|---|
| 704 |
|
|---|
| 705 | # This method is now unused. Using ping_library instead to send the ping action to a
|
|---|
| 706 | # GS2/GS3 server. This method can be used more generally to test whether a URL is alive.
|
|---|
| 707 | # http://search.cpan.org/dist/libwww-perl/lib/LWP/Simple.pm
|
|---|
| 708 | # and http://www.perlmonks.org/?node_id=618534
|
|---|
| 709 | sub is_URL_active {
|
|---|
| 710 | my $url = shift(@_); # gs3 or gs2 URL
|
|---|
| 711 |
|
|---|
| 712 | my $status = 0;
|
|---|
| 713 | if(defined $url) {
|
|---|
| 714 | $ua->timeout(10); # set LWP useragent to 5s max timeout for testing the URL
|
|---|
| 715 | # Need to set this, else it takes I don't know how long to timeout
|
|---|
| 716 | # http://www.perlmonks.org/?node_id=618534
|
|---|
| 717 |
|
|---|
| 718 | $status = LWP::Simple::head($url); # returns empty list of headers if it fails
|
|---|
| 719 | # LWP::Simple::get($url) is more intensive, so don't need to do that
|
|---|
| 720 | #print STDERR "**** $url is alive.\n" if $status;
|
|---|
| 721 | }
|
|---|
| 722 | return $status;
|
|---|
| 723 | }
|
|---|
| 724 |
|
|---|
| 725 | # Pinging seems to always return true, so this method doesn't work
|
|---|
| 726 | sub pingHost {
|
|---|
| 727 | my $url = shift(@_); # gs3 or gs2 URL
|
|---|
| 728 |
|
|---|
| 729 | my $status = 0;
|
|---|
| 730 | if(defined $url) {
|
|---|
| 731 | # Get just the domain. "http://localhost/gsdl?uq=332033495" becomes "localhost"
|
|---|
| 732 | # "http://localhost/greenstone/cgi-bin/library.cgi" becomes "localhost" too
|
|---|
| 733 |
|
|---|
| 734 | #my $host = $url;
|
|---|
| 735 | #$host =~ s@^https?:\/\/(www.)?@@;
|
|---|
| 736 | #$host =~ s@\/.*@@;
|
|---|
| 737 | #print STDERR "**** HOST: $host\n";
|
|---|
| 738 |
|
|---|
| 739 | # More robust way
|
|---|
| 740 | # http://stackoverflow.com/questions/827024/how-do-i-extract-the-domain-out-of-an-url
|
|---|
| 741 | my $uri = URI->new( $url );
|
|---|
| 742 | my $host = $uri->host;
|
|---|
| 743 |
|
|---|
| 744 | # Ping the host. http://perldoc.perl.org/Net/Ping.html
|
|---|
| 745 | my $p = Net::Ping->new();
|
|---|
| 746 | $status = $p->ping($host); # || 0. Appears to set to undef rather than 0
|
|---|
| 747 | print STDERR "**** $host is alive.\n" if $status; #print "$host is alive.\n" if $p->ping($host);
|
|---|
| 748 | $p->close();
|
|---|
| 749 | }
|
|---|
| 750 | # return whether pinging was a success or failure
|
|---|
| 751 | return $status;
|
|---|
| 752 | }
|
|---|
| 753 |
|
|---|
| 754 | 1;
|
|---|
