| 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, $oaiserver_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 | 'oaiserver_name' => $oaiserver_name || "oaiserver",
|
|---|
| 69 | #'gs_mode' => "gs2",
|
|---|
| 70 | 'verbosity' => $verbosity || 2
|
|---|
| 71 | };
|
|---|
| 72 |
|
|---|
| 73 | if ((defined $site) && ($site ne "")) { # GS3
|
|---|
| 74 | $self->{'gs_mode'} = "gs3";
|
|---|
| 75 | } else {
|
|---|
| 76 | $self->{'gs_mode'} = "gs2";
|
|---|
| 77 | }
|
|---|
| 78 |
|
|---|
| 79 | return bless($self, $class);
|
|---|
| 80 | }
|
|---|
| 81 |
|
|---|
| 82 | ## TODO: gsprintf to $self->{'out'} in these 2 print functions
|
|---|
| 83 | ## See buildcolutils.pm new() for setting up $out
|
|---|
| 84 |
|
|---|
| 85 | sub print_task_msg {
|
|---|
| 86 | my $self = shift(@_);
|
|---|
| 87 | my ($task_msg, $verbosity_setting) = @_;
|
|---|
| 88 |
|
|---|
| 89 | $verbosity_setting = $self->{'verbosity'} unless $verbosity_setting;
|
|---|
| 90 | #$verbosity_setting = 1 unless defined $verbosity;
|
|---|
| 91 | if($verbosity_setting >= 1) {
|
|---|
| 92 | print STDERR "\n";
|
|---|
| 93 | print STDERR "************************\n";
|
|---|
| 94 | print STDERR "* $task_msg\n";
|
|---|
| 95 | print STDERR "************************\n";
|
|---|
| 96 | }
|
|---|
| 97 | }
|
|---|
| 98 |
|
|---|
| 99 | # Prints messages if the verbosity is right. Does not add new lines.
|
|---|
| 100 | sub print_msg {
|
|---|
| 101 | my $self = shift(@_);
|
|---|
| 102 | my ($msg, $min_verbosity, $verbosity_setting) = @_;
|
|---|
| 103 |
|
|---|
| 104 | # only display error messages if the current
|
|---|
| 105 | # verbosity setting >= the minimum verbosity level
|
|---|
| 106 | # needed for that message to be displayed.
|
|---|
| 107 |
|
|---|
| 108 | $verbosity_setting = $self->{'verbosity'} unless defined $verbosity_setting;
|
|---|
| 109 | $min_verbosity = 1 unless defined $min_verbosity;
|
|---|
| 110 | if($verbosity_setting >= $min_verbosity) { # by default display all 1 messages
|
|---|
| 111 | print STDERR "$msg";
|
|---|
| 112 | }
|
|---|
| 113 | }
|
|---|
| 114 |
|
|---|
| 115 | # Method to send a command to a GS2 or GS3 library_URL
|
|---|
| 116 | # the commands used in this script can be activate, deactivate, ping,
|
|---|
| 117 | # and is-persistent (is-persistent only implemented for GS2).
|
|---|
| 118 | sub config {
|
|---|
| 119 | my $self = shift(@_);
|
|---|
| 120 | my ($command, $check_message_against_regex, $expected_error_code, $silent, $oai_servlet) = @_;
|
|---|
| 121 |
|
|---|
| 122 | my $library_url = $self->get_library_URL(); #$self->{'library_url'};
|
|---|
| 123 | if($oai_servlet) { # if asked to contact the oaiserver servlet, then
|
|---|
| 124 | # replace the library servlet name with oaiserver servlet name
|
|---|
| 125 | $library_url =~ s@([^/]*?)$@$oai_servlet@;
|
|---|
| 126 | }
|
|---|
| 127 |
|
|---|
| 128 | # Gatherer.java's configGS3Server doesn't use the site variable
|
|---|
| 129 | # so we don't have to either
|
|---|
| 130 |
|
|---|
| 131 | # for GS2, getting the HTTP status isn't enough, we need to read the output
|
|---|
| 132 | # since this is what CollectionManager.config() stipulates.
|
|---|
| 133 | # Using LWP::UserAgent::get($url) for this
|
|---|
| 134 |
|
|---|
| 135 | if(!defined $library_url) {
|
|---|
| 136 | return 0;
|
|---|
| 137 | }
|
|---|
| 138 | else {
|
|---|
| 139 | # ampersands need to be escaped
|
|---|
| 140 | # - with single quotes around it for linux for the cmd to run in bash subshell
|
|---|
| 141 | # - with a ^ before it on windows for the cmd to run in a DOS prompt subshell
|
|---|
| 142 | # - or the entire wget command should be nested in double quotes (single quotes don't work on windows)
|
|---|
| 143 | my $wgetCommand = $command;
|
|---|
| 144 |
|
|---|
| 145 | my $wget_file_path = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "wget");
|
|---|
| 146 | my $tmpfilename = &util::get_tmp_filename(".html"); # random file name with html extension in tmp location in which we'll store the HTML page retrieved by wget
|
|---|
| 147 |
|
|---|
| 148 | # https://www.gnu.org/software/wget/manual/wget.html
|
|---|
| 149 | # output-document set to - (STDOUT), so page is streamed to STDOUT
|
|---|
| 150 | # timeout: 5 seconds, tries: 1
|
|---|
| 151 | # wget sends status information and response code to STDERR, so redirect it to STDOUT
|
|---|
| 152 | # Searching for "perl backtick operator redirect stderr to stdout":
|
|---|
| 153 | # http://www.perlmonks.org/?node=How%20can%20I%20capture%20STDERR%20from%20an%20external%20command%3F
|
|---|
| 154 | ##$wgetCommand = "\"$wget_file_path\" --spider -T 5 -t 1 \"$library_url$wgetCommand\" 2>&1"; # won't save page
|
|---|
| 155 | #$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!!
|
|---|
| 156 | $wgetCommand = "\"$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
|
|---|
| 157 |
|
|---|
| 158 | ##print STDERR "@@@@ $wgetCommand\n";
|
|---|
| 159 |
|
|---|
| 160 | my $response_content = "";
|
|---|
| 161 | my $response_code = undef;
|
|---|
| 162 | #my $response_content = `$wgetCommand`; # Dr Bainbridge advises against using backticks for running a process. If capturing std output, use open():
|
|---|
| 163 | if (open(PIN, "$wgetCommand |")) {
|
|---|
| 164 | while (defined (my $perl_output_line = <PIN>)) {
|
|---|
| 165 | $response_content = $response_content . $perl_output_line;
|
|---|
| 166 | }
|
|---|
| 167 | close(PIN);
|
|---|
| 168 | } else {
|
|---|
| 169 | print STDERR "servercontrol.pm::config() failed to run $wgetCommand\n";
|
|---|
| 170 | }
|
|---|
| 171 |
|
|---|
| 172 |
|
|---|
| 173 | my @lines = split( /\n/, $response_content );
|
|---|
| 174 | foreach my $line (@lines) {
|
|---|
| 175 | #print STDERR "@@@@ LINE: $line\n";
|
|---|
| 176 | if($line =~ m@failed: Connection timed out.$@) { # linux
|
|---|
| 177 | $response_code = "failed: Connection timed out.";
|
|---|
| 178 | last; # break keyword in perl = last
|
|---|
| 179 | }
|
|---|
| 180 | elsif($line =~ m@Giving up.$@) { # windows (unless -T 5 -t 1 is not passed in)
|
|---|
| 181 | $response_code = "failed: Giving up.";
|
|---|
| 182 | last; # break keyword in perl = last
|
|---|
| 183 | }
|
|---|
| 184 | elsif($line =~ m@failed: Connection refused.$@) {
|
|---|
| 185 | $response_code = "failed: Connection refused.";
|
|---|
| 186 | last; # break keyword in perl = last
|
|---|
| 187 | }
|
|---|
| 188 | elsif($line =~ m@HTTP request sent, @) {
|
|---|
| 189 | $response_code = $line;
|
|---|
| 190 | $response_code =~ s@[^\d]*(.*)$@$1@;
|
|---|
| 191 | last;
|
|---|
| 192 | }
|
|---|
| 193 | }
|
|---|
| 194 |
|
|---|
| 195 | if($command =~ m@ping@ && $response_code =~ m@failed: (Connection refused|Giving up)@) {
|
|---|
| 196 | # server not running
|
|---|
| 197 | $self->print_msg("*** Server not running. $library_url$command\n", 3);
|
|---|
| 198 | &FileUtils::removeFiles($tmpfilename); # get rid of the ping response's temporary html file we downloaded
|
|---|
| 199 | return 0;
|
|---|
| 200 | }
|
|---|
| 201 | if($response_code && (index($response_code, "200") != -1)) {
|
|---|
| 202 | $self->print_msg("*** Command $library_url$command\n", 3);
|
|---|
| 203 | $self->print_msg("*** HTTP Response Status: $response_code - Complete.", 3);
|
|---|
| 204 |
|
|---|
| 205 | # check the page content is as expected
|
|---|
| 206 | #my $resultstr = $response_content;
|
|---|
| 207 |
|
|---|
| 208 | # The following file reading section is a candidate to use FileUtils::readUTF8File()
|
|---|
| 209 | # in place of calling sysread() directly. But only if we can reason we'd be working with UTF8
|
|---|
| 210 | open(FIN,"<$tmpfilename") or die "servercontrol.pm: Unable to open $tmpfilename to read ping response page...ERROR: $!\n";
|
|---|
| 211 | my $resultstr;
|
|---|
| 212 | # Read in the entire contents of the file in one hit
|
|---|
| 213 | ##sysread(FIN, $resultstr, -s FIN);
|
|---|
| 214 | # GLI GBuildProgressMonitor.hasSignalledStop() removes any lines that start with elements it doesn't recognise
|
|---|
| 215 | # but we want to see the lines with HTML tags as they are the most interesting: they embed success/failure messages
|
|---|
| 216 | # So read line by line from the downloaded web page and replace <> with [] so that GLI can display these lines too
|
|---|
| 217 | while (my $line = <FIN>) {
|
|---|
| 218 | chomp $line;
|
|---|
| 219 | $line =~ s@<@[@g;
|
|---|
| 220 | $line =~ s@>@]@g;
|
|---|
| 221 | $resultstr .= "$line\n";
|
|---|
| 222 | }
|
|---|
| 223 | close(FIN);
|
|---|
| 224 | &FileUtils::removeFiles($tmpfilename); # get rid of the ping response's temporary html file we downloaded
|
|---|
| 225 |
|
|---|
| 226 |
|
|---|
| 227 | #$resultstr =~ s@.*gs_content\"\>@@s; ## only true for default library servlet
|
|---|
| 228 | #$resultstr =~ s@</div>.*@@s;
|
|---|
| 229 | if($resultstr =~ m/$check_message_against_regex/) {
|
|---|
| 230 | $self->print_msg(" Response as expected.\n", 3);
|
|---|
| 231 | $self->print_msg("@@@@@@ Got result:\n$resultstr\n", 4);
|
|---|
| 232 | return 1;
|
|---|
| 233 | } else {
|
|---|
| 234 | # if we expect the collection to be inactive, then we'd be in silent mode: if so,
|
|---|
| 235 | # don't print out the "ping did not succeed" response, but print out any other messages
|
|---|
| 236 |
|
|---|
| 237 | # So we only suppress the ping col "did not succeed" response if we're in silent mode
|
|---|
| 238 | # But if any message other than ping "did not succeed" is returned, we always print it
|
|---|
| 239 | if($resultstr !~ m/did not succeed/ || !$silent) {
|
|---|
| 240 | $self->print_msg("\n\tBUT: command $library_url$command response UNEXPECTED.\n", 3);
|
|---|
| 241 | $self->print_msg("*** Got message:\n$response_content.\n", 4);
|
|---|
| 242 | $self->print_msg("*** Got result:\n$resultstr\n", 3);
|
|---|
| 243 | }
|
|---|
| 244 | return 0; # ping on a collection may "not succeed."
|
|---|
| 245 | }
|
|---|
| 246 | }
|
|---|
| 247 | elsif($response_code && $response_code =~ m@^(4|5)\d\d@) { # client side errors start with 4xx, server side with 5xx
|
|---|
| 248 | # check the page content is as expected
|
|---|
| 249 | if(defined $expected_error_code && $response_code =~ m@^$expected_error_code@) {
|
|---|
| 250 | $self->print_msg(" Response status $response_code as expected.\n", 3);
|
|---|
| 251 | } else {
|
|---|
| 252 | $self->print_msg("*** Command $library_url$command\n");
|
|---|
| 253 | $self->print_msg("*** Unexpected error type 1. HTTP Response Status: $response_code - Failed.\n");
|
|---|
| 254 | }
|
|---|
| 255 | return 0; # return false, since the response_code was an error, expected or not
|
|---|
| 256 | }
|
|---|
| 257 | else { # also if response_code is still undefined, as can happen with connection timing out
|
|---|
| 258 | $self->print_msg("*** Command $library_url$command\n");
|
|---|
| 259 | if(defined $response_code) {
|
|---|
| 260 | $self->print_msg("*** Unexpected error type 2. HTTP Response Status: $response_code - Failed.\n");
|
|---|
| 261 | } else {
|
|---|
| 262 | $self->print_msg("*** Unexpected error type 3. Failed:\n\n$response_content\n\n");
|
|---|
| 263 | }
|
|---|
| 264 | return 0;
|
|---|
| 265 | }
|
|---|
| 266 | #print STDERR "********** WgetCommand: $wgetCommand\n\n";
|
|---|
| 267 | #print STDERR "********** Response_content:\n$response_content\n\n";
|
|---|
| 268 | #print STDERR "********** Response_CODE: $response_code\n";
|
|---|
| 269 |
|
|---|
| 270 | }
|
|---|
| 271 | }
|
|---|
| 272 |
|
|---|
| 273 | sub deactivate_collection {
|
|---|
| 274 | my $self = shift(@_);
|
|---|
| 275 |
|
|---|
| 276 | my $gs_mode = $self->{'gs_mode'};
|
|---|
| 277 | my $qualified_collection = $self->{'qualified_collection'};
|
|---|
| 278 |
|
|---|
| 279 | if($gs_mode eq "gs2") {
|
|---|
| 280 | my $DEACTIVATE_COMMAND = "?a=config&cmd=release-collection&c=";
|
|---|
| 281 | my $check_message_against_regex = q/configured release-collection/;
|
|---|
| 282 | $self->config($DEACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
|
|---|
| 283 | }
|
|---|
| 284 | elsif ($gs_mode eq "gs3") {
|
|---|
| 285 | $self->print_msg("\t- Main library servlet\n");
|
|---|
| 286 | my $DEACTIVATE_COMMAND = "?excerptid=gs_content&a=s&sa=d&st=collection&sn=";
|
|---|
| 287 | my $check_message_against_regex = "collection: $qualified_collection deactivated";
|
|---|
| 288 | $self->config($DEACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
|
|---|
| 289 |
|
|---|
| 290 | # and deactivate the collection on OAIserver url too.
|
|---|
| 291 | # NOTE: if it's not an OAI collection, then the message that the collection is "not enabled for OAI" is EXPECTED. Another possible valid outcome.
|
|---|
| 292 | $self->print_msg("\t- OAI servlet\n");
|
|---|
| 293 | $DEACTIVATE_COMMAND = "?deactivate=";
|
|---|
| 294 | $check_message_against_regex = "(collection\: $qualified_collection deactivated|collection\: $qualified_collection is not enabled for OAI.)";
|
|---|
| 295 | $self->config($DEACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex, undef, undef, $self->{'oaiserver_name'});
|
|---|
| 296 | }
|
|---|
| 297 | }
|
|---|
| 298 |
|
|---|
| 299 | sub activate_collection {
|
|---|
| 300 | my $self = shift(@_);
|
|---|
| 301 |
|
|---|
| 302 | my $gs_mode = $self->{'gs_mode'};
|
|---|
| 303 | my $qualified_collection = $self->{'qualified_collection'};
|
|---|
| 304 |
|
|---|
| 305 | if($gs_mode eq "gs2") {
|
|---|
| 306 | my $ACTIVATE_COMMAND = "?a=config&cmd=add-collection&c=";
|
|---|
| 307 | my $check_message_against_regex = q/configured add-collection/;
|
|---|
| 308 | $self->config($ACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
|
|---|
| 309 | }
|
|---|
| 310 | elsif ($gs_mode eq "gs3") {
|
|---|
| 311 | $self->print_msg("\t- Main library servlet\n");
|
|---|
| 312 | my $ACTIVATE_COMMAND = "?excerptid=gs_content&a=s&sa=a&st=collection&sn=";
|
|---|
| 313 | my $check_message_against_regex = "collection: $qualified_collection activated";
|
|---|
| 314 | $self->config($ACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex);
|
|---|
| 315 |
|
|---|
| 316 | # and activate the collection on OAIserver url too.
|
|---|
| 317 | # NOTE: if it's not an OAI collection, then the message that the collection is "not enabled for OAI" is EXPECTED. Another possible valid outcome.
|
|---|
| 318 | $self->print_msg("\t- OAI servlet\n");
|
|---|
| 319 | $ACTIVATE_COMMAND = "?activate=";
|
|---|
| 320 | $check_message_against_regex = "(collection\: $qualified_collection activated|collection\: $qualified_collection is not enabled for OAI.)";
|
|---|
| 321 | $self->config($ACTIVATE_COMMAND.$qualified_collection, $check_message_against_regex, undef, undef, $self->{'oaiserver_name'});
|
|---|
| 322 | }
|
|---|
| 323 | }
|
|---|
| 324 |
|
|---|
| 325 | sub ping {
|
|---|
| 326 | my $self = shift(@_);
|
|---|
| 327 | my $command = shift(@_);
|
|---|
| 328 | my $silent = shift(@_);
|
|---|
| 329 |
|
|---|
| 330 | # If the GS server is not running, we *expect* to see a "500" status code.
|
|---|
| 331 | # If the GS server is running, then "Ping" ... "succeeded" is expected on success.
|
|---|
| 332 | # When pinging an inactive collection, it will say it did "not succeed". This is
|
|---|
| 333 | # a message of interest to return.
|
|---|
| 334 | my $check_responsemsg_against_regex = q/(succeeded)/;
|
|---|
| 335 | my $expected_error_code = 500;
|
|---|
| 336 |
|
|---|
| 337 | $self->print_msg("*** COMMAND WAS: |$command| ***\n", 4);
|
|---|
| 338 |
|
|---|
| 339 | return $self->config($command, $check_responsemsg_against_regex, $expected_error_code, $silent);
|
|---|
| 340 | }
|
|---|
| 341 |
|
|---|
| 342 | # send a pingaction to the GS library. General server-level ping.
|
|---|
| 343 | sub ping_library {
|
|---|
| 344 | my $self = shift(@_);
|
|---|
| 345 |
|
|---|
| 346 | my $gs_mode = $self->{'gs_mode'};
|
|---|
| 347 |
|
|---|
| 348 | my $command = "";
|
|---|
| 349 | if($gs_mode eq "gs2") {
|
|---|
| 350 | $command = "?a=ping";
|
|---|
| 351 | }
|
|---|
| 352 | elsif ($gs_mode eq "gs3") {
|
|---|
| 353 | $command = "?a=s&sa=ping&excerptid=gs_content";
|
|---|
| 354 | }
|
|---|
| 355 | return $self->ping($command);
|
|---|
| 356 | }
|
|---|
| 357 |
|
|---|
| 358 |
|
|---|
| 359 | # send a pingaction to a collection in GS library to check if it's active
|
|---|
| 360 | sub ping_library_collection {
|
|---|
| 361 | my $self = shift(@_);
|
|---|
| 362 | my $silent = shift(@_);
|
|---|
| 363 |
|
|---|
| 364 | my $gs_mode = $self->{'gs_mode'};
|
|---|
| 365 | my $qualified_collection = $self->{'qualified_collection'};
|
|---|
| 366 |
|
|---|
| 367 | my $command = "";
|
|---|
| 368 | if($gs_mode eq "gs2") {
|
|---|
| 369 | $command = "?a=ping&c=$qualified_collection";
|
|---|
| 370 | }
|
|---|
| 371 | elsif ($gs_mode eq "gs3") {
|
|---|
| 372 | $command = "?a=s&sa=ping&st=collection&sn=$qualified_collection&excerptid=gs_content";
|
|---|
| 373 | }
|
|---|
| 374 | return $self->ping($command, $silent);
|
|---|
| 375 | }
|
|---|
| 376 |
|
|---|
| 377 | # return true if server is persistent, by calling is-persistent on library_url
|
|---|
| 378 | # this is only for GS2, since the GS3 server is always persistent
|
|---|
| 379 | sub is_persistent {
|
|---|
| 380 | my $self = shift(@_);
|
|---|
| 381 |
|
|---|
| 382 | if($self->{'gs_mode'} eq "gs3") { # GS3 server is always persistent
|
|---|
| 383 | return 1;
|
|---|
| 384 | }
|
|---|
| 385 |
|
|---|
| 386 | my $command = "?a=is-persistent";
|
|---|
| 387 | my $check_responsemsg_against_regex = q/true/; # isPersistent: true versus isPersistent: false
|
|---|
| 388 | return $self->config($command, $check_responsemsg_against_regex);
|
|---|
| 389 | }
|
|---|
| 390 |
|
|---|
| 391 | sub set_library_URL {
|
|---|
| 392 | my $self = shift(@_);
|
|---|
| 393 | my $library_url = shift(@_);
|
|---|
| 394 | $self->{'library_url'} = $library_url;
|
|---|
| 395 | }
|
|---|
| 396 |
|
|---|
| 397 | sub get_library_URL {
|
|---|
| 398 | my $self = shift(@_);
|
|---|
| 399 |
|
|---|
| 400 | # For web servers that are external to a Greenstone installation,
|
|---|
| 401 | # the user can pass in their web server's library URL.
|
|---|
| 402 | if($self->{'library_url'}) {
|
|---|
| 403 | return $self->{'library_url'};
|
|---|
| 404 | }
|
|---|
| 405 |
|
|---|
| 406 | # For web servers included with GS (like tomcat for GS3 and server.exe
|
|---|
| 407 | # and apache for GS2), we work out the library URL:
|
|---|
| 408 | my ($gs_mode, $lib_name); # gs_mode can be gs3 or gs2, lib_name is the custom servlet name
|
|---|
| 409 | $gs_mode = $self->{'gs_mode'};
|
|---|
| 410 | $lib_name = $self->{'library_name'};
|
|---|
| 411 |
|
|---|
| 412 | # If we get here, we are dealing with a server included with GS.
|
|---|
| 413 | # For GS3, we ask ant for the library URL.
|
|---|
| 414 | # For GS2, we derive the URL from the llssite.cfg file.
|
|---|
| 415 |
|
|---|
| 416 | # note that unless we pass in $get_public_url=1, we now get the local http URL
|
|---|
| 417 | # by default (e.g. http://127.0.0.1:httpPort/greenstone/library)
|
|---|
| 418 | my $url = &util::get_full_greenstone_url_prefix($gs_mode, $lib_name); # found largely identical method copied
|
|---|
| 419 | # into util.pm. Don't want duplicates, so calling that from here.
|
|---|
| 420 |
|
|---|
| 421 | # either the url is still undef or it is now set
|
|---|
| 422 | #print STDERR "\n@@@@@ final URL:|$url|\n" if $url;
|
|---|
| 423 | #print STDERR "\n@@@@@ URL still undef\n" if !$url;
|
|---|
| 424 |
|
|---|
| 425 | if (defined $url) {
|
|---|
| 426 | $self->{'library_url'} = $url;
|
|---|
| 427 | }
|
|---|
| 428 |
|
|---|
| 429 | return $url;
|
|---|
| 430 | }
|
|---|
| 431 |
|
|---|
| 432 | sub do_deactivate {
|
|---|
| 433 | my $self = shift(@_);
|
|---|
| 434 |
|
|---|
| 435 | # 1. Get library URL
|
|---|
| 436 |
|
|---|
| 437 | # For web servers that are external to a Greenstone installation,
|
|---|
| 438 | # the user can pass in their web server's library URL.
|
|---|
| 439 | # For web servers included with GS (like tomcat for GS3 and server.exe
|
|---|
| 440 | # and apache for GS2), we work out the library URL:
|
|---|
| 441 |
|
|---|
| 442 | # Can't do $self->{'library_url'}, because it may not yet be set
|
|---|
| 443 | my $library_url = $self->get_library_URL(); # returns undef if no valid server URL
|
|---|
| 444 |
|
|---|
| 445 | if(!defined $library_url) { # undef if no valid server URL
|
|---|
| 446 | return; # can't do any deactivation without a valid server URL
|
|---|
| 447 | }
|
|---|
| 448 |
|
|---|
| 449 | my $is_persistent_server = $self->{'is_persistent_server'};
|
|---|
| 450 | my $qualified_collection = $self->{'qualified_collection'};
|
|---|
| 451 |
|
|---|
| 452 | # CollectionManager's installCollection phase in GLI
|
|---|
| 453 | # 2. Ping the library URL, and if it's a persistent server and running, release the collection
|
|---|
| 454 |
|
|---|
| 455 | $self->print_msg("Pinging $library_url\n");
|
|---|
| 456 | if ($self->ping_library()) { # server running
|
|---|
| 457 |
|
|---|
| 458 | # server is running, so release the collection if
|
|---|
| 459 | # the server is persistent and the collection is active
|
|---|
| 460 |
|
|---|
| 461 | # don't need to work out persistency of server more than once, since the libraryURL hasn't changed
|
|---|
| 462 | if (!defined $is_persistent_server) {
|
|---|
| 463 | $self->print_msg("Checking if Greenstone server is persistent\n");
|
|---|
| 464 | $is_persistent_server = $self->is_persistent();
|
|---|
| 465 | $self->{'is_persistent_server'} = $is_persistent_server;
|
|---|
| 466 | }
|
|---|
| 467 |
|
|---|
| 468 | if ($is_persistent_server) { # only makes sense to issue activate and deactivate cmds to a persistent server
|
|---|
| 469 |
|
|---|
| 470 | $self->print_msg("Checking if the collection $qualified_collection is already active\n");
|
|---|
| 471 | my $collection_active = $self->ping_library_collection();
|
|---|
| 472 |
|
|---|
| 473 | if ($collection_active) {
|
|---|
| 474 | $self->print_msg("De-activating collection $qualified_collection\n");
|
|---|
| 475 | $self->deactivate_collection();
|
|---|
| 476 | }
|
|---|
| 477 | else {
|
|---|
| 478 | $self->print_msg("Collection is not active => No need to deactivate\n");
|
|---|
| 479 | }
|
|---|
| 480 | }
|
|---|
| 481 | else {
|
|---|
| 482 | $self->print_msg("Server is not persistent => No need to deactivate collection\n");
|
|---|
| 483 | }
|
|---|
| 484 | }
|
|---|
| 485 | else {
|
|---|
| 486 | $self->print_msg("No response to Ping => Taken to mean server is not running\n");
|
|---|
| 487 | }
|
|---|
| 488 |
|
|---|
| 489 | return $is_persistent_server;
|
|---|
| 490 | }
|
|---|
| 491 |
|
|---|
| 492 | sub do_activate {
|
|---|
| 493 | my $self = shift @_;
|
|---|
| 494 |
|
|---|
| 495 | my $library_url = $self->get_library_URL(); # Can't do $self->{'library_url'}; as it may not be set yet
|
|---|
| 496 |
|
|---|
| 497 | if(!defined $library_url) { # undef if no valid server URL
|
|---|
| 498 | return; # nothing to activate if without valid server URL
|
|---|
| 499 | }
|
|---|
| 500 |
|
|---|
| 501 | my $is_persistent_server = $self->{'is_persistent_server'};
|
|---|
| 502 | my $qualified_collection = $self->{'qualified_collection'};
|
|---|
| 503 |
|
|---|
| 504 | $self->print_msg("Pinging $library_url\n");
|
|---|
| 505 | if ($self->ping_library()) { # server running
|
|---|
| 506 |
|
|---|
| 507 | # don't need to work out persistency of server more than once, since the libraryURL hasn't changed
|
|---|
| 508 | if (!defined $is_persistent_server) {
|
|---|
| 509 | $self->print_msg("Checking if Greenstone server is persistent\n");
|
|---|
| 510 | $is_persistent_server = $self->is_persistent();
|
|---|
| 511 | $self->{'is_persistent_server'} = $is_persistent_server;
|
|---|
| 512 | }
|
|---|
| 513 |
|
|---|
| 514 | if ($is_persistent_server) { # persistent server, so can try activating collection
|
|---|
| 515 |
|
|---|
| 516 | $self->print_msg("Checking if the collection $qualified_collection is not already active\n");
|
|---|
| 517 |
|
|---|
| 518 | # Since we could have deactivated the collection at this point,
|
|---|
| 519 | # it is likely that it is not yet active. When pinging the collection
|
|---|
| 520 | # a "ping did not succeed" message is expected, therefore tell the ping
|
|---|
| 521 | # to proceed silently
|
|---|
| 522 | my $silent = 1;
|
|---|
| 523 | my $collection_active = $self->ping_library_collection($silent);
|
|---|
| 524 |
|
|---|
| 525 | if (!$collection_active) {
|
|---|
| 526 | $self->print_msg(" Collection is not active.\n");
|
|---|
| 527 | $self->print_msg("Activating collection $qualified_collection\n");
|
|---|
| 528 | $self->activate_collection();
|
|---|
| 529 |
|
|---|
| 530 | # unless an error occurred, the collection should now be active:
|
|---|
| 531 | $collection_active = $self->ping_library_collection(); # not silent if ping did not succeed
|
|---|
| 532 | if(!$collection_active) {
|
|---|
| 533 | $self->print_msg("ERROR: collection $qualified_collection did not get activated\n");
|
|---|
| 534 | $self->print_msg(" (if the collection's not in the default library, you must pass\n");
|
|---|
| 535 | $self->print_msg(" in the -library_name flag with the correct name of the library.)\n");
|
|---|
| 536 | }
|
|---|
| 537 | }
|
|---|
| 538 | else {
|
|---|
| 539 | $self->print_msg("Collection is already active => No need to activate\n");
|
|---|
| 540 | }
|
|---|
| 541 | }
|
|---|
| 542 | else {
|
|---|
| 543 | $self->print_msg("Server is not persistent => No need to activate collection\n");
|
|---|
| 544 | }
|
|---|
| 545 | }
|
|---|
| 546 | else {
|
|---|
| 547 | $self->print_msg("No response to Ping => Taken to mean server is not running\n");
|
|---|
| 548 | }
|
|---|
| 549 |
|
|---|
| 550 | return $is_persistent_server;
|
|---|
| 551 | }
|
|---|
| 552 |
|
|---|
| 553 |
|
|---|
| 554 | #########################################################
|
|---|
| 555 | ### UNUSED METHODS - CAN BE HANDY
|
|---|
| 556 |
|
|---|
| 557 |
|
|---|
| 558 | # This method uses the perl libraries we're advised to use in place of wget for pinging and retrieving web
|
|---|
| 559 | # pages. The problem is that not all perl installations may support these libraries. So we now use the new
|
|---|
| 560 | # config() method further above, which uses the wget included in Greenstone binary installations.
|
|---|
| 561 | # If the library imports at page top conflict, comment out those imports and move the methods config_old(),
|
|---|
| 562 | # is_URL_active() and pingHost() out to a temporary file.
|
|---|
| 563 | #
|
|---|
| 564 | # If for some reason you can't use wget, then rename the config() method to config_old(), and rename the
|
|---|
| 565 | # method below to config() and things should work as before.
|
|---|
| 566 | sub config_old {
|
|---|
| 567 | my $self = shift(@_);
|
|---|
| 568 | my ($command, $check_message_against_regex, $expected_error_code, $silent, $oai_servlet) = @_;
|
|---|
| 569 |
|
|---|
| 570 | my $library_url = $self->get_library_URL(); #$self->{'library_url'};
|
|---|
| 571 | if($oai_servlet) { # if asked to contact the oaiserver servlet, then
|
|---|
| 572 | # replace the library servlet name with oaiserver servlet name
|
|---|
| 573 | $library_url =~ s@([^/]*?)$@$oai_servlet@;
|
|---|
| 574 | }
|
|---|
| 575 |
|
|---|
| 576 | # Gatherer.java's configGS3Server doesn't use the site variable
|
|---|
| 577 | # so we don't have to either
|
|---|
| 578 |
|
|---|
| 579 | # for GS2, getting the HTTP status isn't enough, we need to read the output
|
|---|
| 580 | # since this is what CollectionManager.config() stipulates.
|
|---|
| 581 | # Using LWP::UserAgent::get($url) for this
|
|---|
| 582 |
|
|---|
| 583 | if(!defined $library_url) {
|
|---|
| 584 | return 0;
|
|---|
| 585 | }
|
|---|
| 586 | else {
|
|---|
| 587 | $ua->timeout(5); # set LWP useragent to 5s max timeout for testing the URL
|
|---|
| 588 | # Need to set this, else it takes I don't know how long to timeout
|
|---|
| 589 | # http://www.perlmonks.org/?node_id=618534
|
|---|
| 590 |
|
|---|
| 591 | # http://search.cpan.org/~gaas/libwww-perl-6.04/lib/LWP/UserAgent.pm
|
|---|
| 592 | # use LWP::UserAgent's get($url) since it returns an HTTP::Response code
|
|---|
| 593 |
|
|---|
| 594 | my $response_obj = $ua->get($library_url.$command);
|
|---|
| 595 |
|
|---|
| 596 | # $response_obj->content stores the content and $response_obj->code the HTTP response code
|
|---|
| 597 | my $response_code = $response_obj->code();
|
|---|
| 598 |
|
|---|
| 599 | if(LWP::Simple::is_success($response_code)) {# $response_code eq RC_OK) { # LWP::Simple::is_success($response_code)
|
|---|
| 600 | $self->print_msg("*** Command $library_url$command\n", 3);
|
|---|
| 601 | $self->print_msg("*** HTTP Response Status: $response_code - Complete.", 3);
|
|---|
| 602 |
|
|---|
| 603 | # check the page content is as expected
|
|---|
| 604 | my $response_content = $response_obj->content;
|
|---|
| 605 | my $resultstr = $response_content;
|
|---|
| 606 | $resultstr =~ s@.*gs_content\"\>@@s;
|
|---|
| 607 | $resultstr =~ s@</div>.*@@s;
|
|---|
| 608 |
|
|---|
| 609 | if($resultstr =~ m/$check_message_against_regex/) {#if($response_content =~ m/$check_message_against_regex/) {
|
|---|
| 610 | $self->print_msg(" Response as expected.\n", 3);
|
|---|
| 611 | $self->print_msg("@@@@@@ Got result:\n$resultstr\n", 4);
|
|---|
| 612 | return 1;
|
|---|
| 613 | } else {
|
|---|
| 614 | # if we expect the collection to be inactive, then we'd be in silent mode: if so,
|
|---|
| 615 | # don't print out the "ping did not succeed" response, but print out any other messages
|
|---|
| 616 |
|
|---|
| 617 | # So we only suppress the ping col "did not succeed" response if we're in silent mode
|
|---|
| 618 | # But if any message other than ping "did not succeed" is returned, we always print it
|
|---|
| 619 | if($resultstr !~ m/did not succeed/ || !$silent) {#if($response_content !~ m/did not succeed/ || !$silent) {
|
|---|
| 620 | $self->print_msg("\n\tBUT: command $library_url$command response UNEXPECTED.\n", 3);
|
|---|
| 621 | $self->print_msg("*** Got message:\n$response_content.\n", 4);
|
|---|
| 622 | $self->print_msg("*** Got result:\n$resultstr\n", 3);
|
|---|
| 623 | }
|
|---|
| 624 | return 0; # ping on a collection may "not succeed."
|
|---|
| 625 | }
|
|---|
| 626 | }
|
|---|
| 627 | elsif(LWP::Simple::is_error($response_code)) { # method exported by LWP::Simple, along with HTTP::Status constants
|
|---|
| 628 | # check the page content is as expected
|
|---|
| 629 | if(defined $expected_error_code && $response_code == $expected_error_code) {
|
|---|
| 630 | $self->print_msg(" Response status $response_code as expected.\n", 3);
|
|---|
| 631 | } else {
|
|---|
| 632 | $self->print_msg("*** Command $library_url$command\n");
|
|---|
| 633 | $self->print_msg("*** Unexpected error. HTTP Response Status: $response_code - Failed.\n");
|
|---|
| 634 | }
|
|---|
| 635 | return 0; # return false, since the response_code was an error, expected or not
|
|---|
| 636 | }
|
|---|
| 637 | else {
|
|---|
| 638 | $self->print_msg("*** Command $library_url$command\n");
|
|---|
| 639 | $self->print_msg("*** Unexpected error. HTTP Response Status: $response_code - Failed.\n");
|
|---|
| 640 | return 0;
|
|---|
| 641 | }
|
|---|
| 642 | }
|
|---|
| 643 | }
|
|---|
| 644 |
|
|---|
| 645 | # This method is now unused. Using ping_library instead to send the ping action to a
|
|---|
| 646 | # GS2/GS3 server. This method can be used more generally to test whether a URL is alive.
|
|---|
| 647 | # http://search.cpan.org/dist/libwww-perl/lib/LWP/Simple.pm
|
|---|
| 648 | # and http://www.perlmonks.org/?node_id=618534
|
|---|
| 649 | sub is_URL_active {
|
|---|
| 650 | my $url = shift(@_); # gs3 or gs2 URL
|
|---|
| 651 |
|
|---|
| 652 | my $status = 0;
|
|---|
| 653 | if(defined $url) {
|
|---|
| 654 | $ua->timeout(10); # set LWP useragent to 5s max timeout for testing the URL
|
|---|
| 655 | # Need to set this, else it takes I don't know how long to timeout
|
|---|
| 656 | # http://www.perlmonks.org/?node_id=618534
|
|---|
| 657 |
|
|---|
| 658 | $status = LWP::Simple::head($url); # returns empty list of headers if it fails
|
|---|
| 659 | # LWP::Simple::get($url) is more intensive, so don't need to do that
|
|---|
| 660 | #print STDERR "**** $url is alive.\n" if $status;
|
|---|
| 661 | }
|
|---|
| 662 | return $status;
|
|---|
| 663 | }
|
|---|
| 664 |
|
|---|
| 665 | # Pinging seems to always return true, so this method doesn't work
|
|---|
| 666 | sub pingHost {
|
|---|
| 667 | my $url = shift(@_); # gs3 or gs2 URL
|
|---|
| 668 |
|
|---|
| 669 | my $status = 0;
|
|---|
| 670 | if(defined $url) {
|
|---|
| 671 | # Get just the domain. "http://localhost/gsdl?uq=332033495" becomes "localhost"
|
|---|
| 672 | # "http://localhost/greenstone/cgi-bin/library.cgi" becomes "localhost" too
|
|---|
| 673 |
|
|---|
| 674 | #my $host = $url;
|
|---|
| 675 | #$host =~ s@^https?:\/\/(www.)?@@;
|
|---|
| 676 | #$host =~ s@\/.*@@;
|
|---|
| 677 | #print STDERR "**** HOST: $host\n";
|
|---|
| 678 |
|
|---|
| 679 | # More robust way
|
|---|
| 680 | # http://stackoverflow.com/questions/827024/how-do-i-extract-the-domain-out-of-an-url
|
|---|
| 681 | my $uri = URI->new( $url );
|
|---|
| 682 | my $host = $uri->host;
|
|---|
| 683 |
|
|---|
| 684 | # Ping the host. http://perldoc.perl.org/Net/Ping.html
|
|---|
| 685 | my $p = Net::Ping->new();
|
|---|
| 686 | $status = $p->ping($host); # || 0. Appears to set to undef rather than 0
|
|---|
| 687 | print STDERR "**** $host is alive.\n" if $status; #print "$host is alive.\n" if $p->ping($host);
|
|---|
| 688 | $p->close();
|
|---|
| 689 | }
|
|---|
| 690 | # return whether pinging was a success or failure
|
|---|
| 691 | return $status;
|
|---|
| 692 | }
|
|---|
| 693 |
|
|---|
| 694 | 1;
|
|---|
