modxcms
Blank page at sub.domain.com/manager/
Bug report
Summary
After "Installation finished in 1.9130 s" following instructions, I see only a blank page at /manager/. I'm hosting the page with nginx.
Step to reproduce
wget https://modx.com/download/direct/modx-3.0.1-pl-sdk.zip
unzip modx-3.0.1-pl.zip
mv modx-3.0.1-pl sub.domain.com
cd sub.domain.com/setup/
Vim replacements after I copy config.dist.new.xml to config.xml:
:%s@/www/modx/@/var/www/sub.domain.com@gc
:%s/db_username/root/gc
:%s/db_password/pass/gc
:%s@/modx/@/@gc
I change <http_host>localhost</http_host> to <http_host>sub.domain.com</http_host>
I run:
php ./index.php --installmode=new --core_path=/var/www/sub.domain.com/core/
My /etc/php/8.0/fpm/php.ini has this:
pdo_mysql.default_socket=/var/run/mysqld/mysqlx.sock
mysqli.default_socket = /var/run/mysqld/mysqlx.sock
I ensure the fastcgi_pass in my nginx config is
fastcgi_pass unix:/run/php/php-fpm.sock;
which matches my /etc/php/8.0/fpm/pool.d/www.conf :
listen = /run/php/php-fpm.sock
This code runs correctly:
<?php
/* Connect to an ODBC database using driver invocation */
$dsn = 'mysql:dbname=testdb;host=localhost';
$user = 'dbuser';
$password = 'dbpass';
try {
$dbh = new PDO($dsn, $user, $password);
} catch (PDOException $e) {
echo 'Connection failed: ' . $e->getMessage();
}
?>
Observed behavior
The manager page is blank. It is a 200 response with a completely blank file. Here are the response headers:
Cache-Control: no-store, no-cache, must-revalidate
Connection: keep-alive
Content-Encoding: gzip
Content-Type: text/html; charset=UTF-8
Date: Fri, 22 Jul 2022 20:17:52 GMT
Expires: Thu, 19 Nov 1981 08:52:00 GMT
Pragma: no-cache
Server: nginx/1.18.0 (Ubuntu)
Transfer-Encoding: chunked
In my nginx logs, i see the following:
2022/07/22 19:55:16 [error] 17007#17007: *7 FastCGI sent in stderr: "PHP message: PHP Warning: Trying to access array offset on value of type bool in /var/www/sub.domain.com/core/vendor/xpdo/xpdo/src/xPDO/Om/xPDOObject.php on line 251PHP message: PHP Warning: Trying to access array offset on value of type bool in /var/www/sub.domain.com/core/vendor/xpdo/xpdo/src/xPDO/Om/xPDOObject.php on line 251" while reading response header from upstream, client: 71.232.26.159, server: sub.domain.com, request: "GET /manager/ HTTP/1.1", upstream: "fastcgi://unix:/run/php/php-fpm.sock:", host: "sub.domain.com"
Expected behavior
The manager console should be shown... right?
Environment
My Versions
ModX 3.01 PHP 8.0.21 Nginx 1.18.0 Ubuntu 20.04.4 LTS mysql Ver 14.14 Distrib 5.7.37, for Linux Google Chrome browser, for MacOS
My Nginx config
server {
server_name sub.domain.com;
root /var/www/sub.domain.com;
index index.php;
client_max_body_size 30M;
listen [::]:443 ssl; # managed by Certbot
listen 443 ssl; # managed by Certbot
location / {
root /var/www/sub.domain.com;
if (!-e $request_filename) {
rewrite ^/(.*)$ /index.php?q=$1 last;
}
}
location ~ \.php$ {
try_files $uri =404;
fastcgi_split_path_info ^(.+\.php)(.*)$;
fastcgi_pass unix:/run/php/php-fpm.sock;
fastcgi_index index.php;
fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
include fastcgi_params;
fastcgi_ignore_client_abort on;
fastcgi_param SERVER_NAME $http_host;
}
location ~ /\.ht {
deny all;
}
ssl_certificate /etc/letsencrypt/live/sub.domain.com/fullchain.pem; # managed by Certbot
ssl_certificate_key /etc/letsencrypt/live/sub.domain.com/privkey.pem; # managed by Certbot
include /etc/letsencrypt/options-ssl-nginx.conf; # managed by Certbot
ssl_dhparam /etc/letsencrypt/ssl-dhparams.pem; # managed by Certbot
}
My core/config.inc.php:
<?php^M
/**^M
* MODX Configuration file^M
*/^M
$database_type = 'mysql';^M
$database_server = 'localhost';^M
$database_user = 'root';^M
$database_password = '************';^M
$database_connection_charset = 'utf8';^M
$dbase = 'modx_modx';^M
$table_prefix = 'modx_';^M
$database_dsn = 'mysql:host=localhost;dbname=modx_modx;charset=utf8';^M
$config_options = array (
'override_table' => 'MyISAM',
);^M
$driver_options = array (
);^M
^M
$lastInstallTime = 1658519707;^M
^M
$site_id = 'modx62db009b3a1610.81763680';^M
$site_sessionname = 'SN62db009b3376f';^M
$https_port = '443';^M
$uuid = 'c56fa809-53c7-407a-90b7-16139bb1cbef';^M
^M
if (!defined('MODX_CORE_PATH')) {^M
$modx_core_path= '/var/www/domain.example.com/core/';^M
define('MODX_CORE_PATH', $modx_core_path);^M
}^M
if (!defined('MODX_PROCESSORS_PATH')) {^M
$modx_processors_path= '/var/www/domain.example.com/core/src/Revolution/Processors/';^M
define('MODX_PROCESSORS_PATH', $modx_processors_path);^M
}^M
if (!defined('MODX_CONNECTORS_PATH')) {^M
$modx_connectors_path= '/var/www/domain.example.com/connectors/';^M
$modx_connectors_url= '/connectors/';^M
define('MODX_CONNECTORS_PATH', $modx_connectors_path);^M
define('MODX_CONNECTORS_URL', $modx_connectors_url);^M
}^M
if (!defined('MODX_MANAGER_PATH')) {^M
$modx_manager_path= '/var/www/domain.example.com/manager/';^M
$modx_manager_url= '/manager/';^M
define('MODX_MANAGER_PATH', $modx_manager_path);^M
define('MODX_MANAGER_URL', $modx_manager_url);^M
}^M
if (!defined('MODX_BASE_PATH')) {^M
$modx_base_path= '/var/www/domain.example.com';^M
$modx_base_url= '/';^M
define('MODX_BASE_PATH', $modx_base_path);^M
define('MODX_BASE_URL', $modx_base_url);^M
}^M
if(defined('PHP_SAPI') && (PHP_SAPI == "cli" || PHP_SAPI == "embed")) {^M
$isSecureRequest = false;^M
} else {^M
$isSecureRequest = ((isset($_SERVER['HTTPS']) && !empty($_SERVER['HTTPS']) && strtolower($_SERVER['HTTPS']) !== 'off') || $_SERVER['SERVER_PORT'] == $https_port);^M
}^M
if (!defined('MODX_URL_SCHEME')) {^M
$url_scheme= $isSecureRequest ? 'https://' : 'http://';^M
define('MODX_URL_SCHEME', $url_scheme);^M
}^M
if (!defined('MODX_HTTP_HOST')) {^M
if(defined('PHP_SAPI') && (PHP_SAPI == "cli" || PHP_SAPI == "embed")) {^M
$http_host='domain.example.com';^M
define('MODX_HTTP_HOST', $http_host);^M
} else {^M
$http_host= array_key_exists('HTTP_HOST', $_SERVER) ? htmlspecialchars($_SERVER['HTTP_HOST'], ENT_QUOTES) : 'domain.example.com';^M
if ($_SERVER['SERVER_PORT'] != 80) {^M
$http_host = str_replace(':' . $_SERVER['SERVER_PORT'], '', $http_host); // remove port from HTTP_HOST^M
}^M
$http_host .= in_array($_SERVER['SERVER_PORT'], [80, 443]) ? '' : ':' . $_SERVER['SERVER_PORT'];^M
define('MODX_HTTP_HOST', $http_host);^M
}^M
}^M
if (!defined('MODX_SITE_URL')) {^M
$site_url= $url_scheme . $http_host . MODX_BASE_URL;^M
define('MODX_SITE_URL', $site_url);^M
}^M
if (!defined('MODX_ASSETS_PATH')) {^M
$modx_assets_path= '/var/www/domain.example.com/assets/';^M
$modx_assets_url= '/assets/';^M
define('MODX_ASSETS_PATH', $modx_assets_path);^M
define('MODX_ASSETS_URL', $modx_assets_url);^M
}^M
if (!defined('MODX_LOG_LEVEL_FATAL')) {^M
define('MODX_LOG_LEVEL_FATAL', 0);^M
define('MODX_LOG_LEVEL_ERROR', 1);^M
define('MODX_LOG_LEVEL_WARN', 2);^M
define('MODX_LOG_LEVEL_INFO', 3);^M
define('MODX_LOG_LEVEL_DEBUG', 4);^M
}^M
~
Any guidance at all would be helpful!
In my nginx logs, something looks a bit off with the number of colons/protocols in this "upstream" request:
upstream: "fastcgi://unix:/run/php/php-fpm.sock:"
Are all those ^M in the actual config file? That's kinda weird looking.
MODX doesn't control that nginx upstream configuration, I don't know enough about nginx to really give guidance on that.
Is there anything in the browser dev tools when you try to access the manager? And there's absolutely nothing in the response to the 200 response?
What about the MODX log, at core/cache/logs/error.log? Anything there?
There should also be a setup log in core/cache/logs/ (with the timestamp in the filename), worth reviewing that for any warnings or errors.
The PHP warnings in your log I believe are resolved in 3.0.2 with an xPDO update.
Thank you so much for your reply!
^M is how vim displays 0xD, the carriage return character. It shows up when people or software use both types of return characters.
Nothing shows up in the browser dev tools expect for the 200 network request, which returns a fully empty file of 0 bytes. The total size is 344 bytes, which is just the headers.
Here is the MODX log:
[2022-07-22 19:55:09] (ERROR @ /var/www/domain.example.com/setup/includes/modinstallsettings.class.php : 164) PHP warning: fopen(/var/www/domain.example.com/core/cache/setup/settings.cache.php): Failed to open stream: No such file or directory
Oh, here's something obvious in the setup log...
[2022-07-22 19:55:07] (ERROR in xPDO\xPDOConnection::connect @ /var/www/domain.example.com/core/vendor/xpdo/xpdo/src/xPDO/xPDOConnection.php : 89) SQLSTATE[HY000] [1049] Unknown database 'modx_modx'
[2022-07-22 19:55:07] (ERROR in xPDO\xPDOConnection::connect @ /var/www/domain.example.com/core/vendor/xpdo/xpdo/src/xPDO/xPDOConnection.php : 89) SQLSTATE[HY000] [1049] Unknown database 'modx_modx'
[2022-07-22 19:55:07] (ERROR in xPDO\xPDO::getValue @ /var/www/domain.example.com/core/vendor/xpdo/xpdo/src/xPDO/xPDO.php : 1111) No valid PDOStatement provided to getValue
[2022-07-22 19:55:07] (ERROR in xPDO\xPDOConnection::connect @ /var/www/domain.example.com/core/vendor/xpdo/xpdo/src/xPDO/xPDOConnection.php : 89) SQLSTATE[HY000] [1049] Unknown database 'modx_modx'
When I login to mysql command line and type:
use modx_modx;
mysql> show tables;
I see that the DB does in fact appear to have 70 tables:
+------------------------------------+
| Tables_in_modx_modx |
+------------------------------------+
| modx_access_actiondom |
| modx_access_category |
| modx_access_context |
| modx_access_elements |
| modx_access_media_source |
| modx_access_menus |
| modx_access_namespace |
| modx_access_permissions |
| modx_access_policies |
| modx_access_policy_template_groups |
| modx_access_policy_templates |
| modx_access_resource_groups |
| modx_access_resources |
| modx_access_templatevars |
| modx_actiondom |
| modx_actions_fields |
| modx_active_users |
| modx_categories |
| modx_categories_closure |
| modx_content_type |
| modx_context |
| modx_context_resource |
| modx_context_setting |
| modx_dashboard |
| modx_dashboard_widget |
| modx_dashboard_widget_placement |
| modx_deprecated_call |
| modx_deprecated_method |
| modx_document_groups |
| modx_documentgroup_names |
| modx_element_property_sets |
| modx_extension_packages |
| modx_fc_profiles |
| modx_fc_profiles_usergroups |
| modx_fc_sets |
| modx_lexicon_entries |
| modx_manager_log |
| modx_media_sources |
| modx_media_sources_contexts |
| modx_media_sources_elements |
| modx_member_groups |
| modx_membergroup_names |
| modx_menus |
| modx_namespaces |
| modx_property_set |
| modx_register_messages |
| modx_register_queues |
| modx_register_topics |
| modx_session |
| modx_site_content |
| modx_site_htmlsnippets |
| modx_site_plugin_events |
| modx_site_plugins |
| modx_site_snippets |
| modx_site_templates |
| modx_site_tmplvar_access |
| modx_site_tmplvar_contentvalues |
| modx_site_tmplvar_templates |
| modx_site_tmplvars |
| modx_system_eventnames |
| modx_system_settings |
| modx_transport_packages |
| modx_transport_providers |
| modx_user_attributes |
| modx_user_group_roles |
| modx_user_group_settings |
| modx_user_messages |
| modx_user_settings |
| modx_users |
| modx_workspaces |
+------------------------------------+
70 rows in set (0.00 sec)
In your earlier post it looks like you're using root, and setup apparently managed to create tables... but I'd almost think the user may not have access to the specific database?
Perhaps with the installation you relied on the CLI user having access, while the front-end PHP accesses MySQL differently and hasn't been authorized on whatever that internal IP/definition may be?
I'm doing everything as root, and I've specified that MODX should use the root user and password. See the core/config.inc.php:
$database_user = 'root';^M
$database_password = '************';^M
Reading your second sentence, that does seem likely. I can try setting up a non-root db user. I'm off for the weekend on this, but I'll report back next week. It seems like you may have the answer though.
I’ve created a new MySQL user with ALL PRIVILEGES, and I updated my core/config/config.inc.php with the new username and password. I’lll keep trying things. Maybe the setup needs to be run with the new user credenrials?
The issue seems to be with fastcgi and PHP
I will update if I find a solution— then I’d make a PR against the modx nginx docs if I find the right changes to my nginx config.
I followed nginx fastcgi instructions to some success.
My nginx config location looks like this now:
location ~ [^/]\.php(/|$) {
fastcgi_split_path_info ^(.+?\.php)(/.*)$;
if (!-f $document_root$fastcgi_script_name) {
return 404;
}
# Mitigate https://httpoxy.org/ vulnerabilities
fastcgi_param HTTP_PROXY "";
fastcgi_pass unix:/run/php/php-fpm.sock;
fastcgi_index index.php;
include fastcgi.conf;
}
I replaced manager/index.php with this:
<?php var_export($_SERVER)?>
and I see what seems like a valid result on visiting domain.example.com/manager, so it seems fastcgi and nginx are working as advertised:
array ( 'USER' => 'www-data', 'HOME' => '/var/www', 'HTTP_CONNECTION' => 'keep-alive', 'HTTP_ACCEPT_ENCODING' => 'gzip, deflate, br', 'HTTP_ACCEPT_LANGUAGE' => 'en-US,en;q=0.9', 'HTTP_USER_AGENT' => 'Mozilla/5.0 (iPhone; CPU iPhone OS 15_5 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/15.5 Mobile/15E148 Safari/604.1', 'HTTP_COOKIE' => 'PHPSESSID=******', 'HTTP_ACCEPT' => 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8', 'HTTP_HOST' => 'domain.example.com', 'REDIRECT_STATUS' => '200', 'SERVER_NAME' => 'domain.example.com', 'SERVER_PORT' => '443', 'SERVER_ADDR' => '****', 'REMOTE_USER' => '', 'REMOTE_PORT' => '57927', 'REMOTE_ADDR' => '*****', 'SERVER_SOFTWARE' => 'nginx/1.18.0', 'GATEWAY_INTERFACE' => 'CGI/1.1', 'HTTPS' => 'on', 'REQUEST_SCHEME' => 'https', 'SERVER_PROTOCOL' => 'HTTP/1.1', 'DOCUMENT_ROOT' => '/var/www/domain.example.com', 'DOCUMENT_URI' => '/manager/index.php', 'REQUEST_URI' => '/manager/', 'SCRIPT_NAME' => '/manager/index.php', 'CONTENT_LENGTH' => '', 'CONTENT_TYPE' => '', 'REQUEST_METHOD' => 'GET', 'QUERY_STRING' => '', 'SCRIPT_FILENAME' => '/var/www/domain.example.com/manager/index.php', 'FCGI_ROLE' => 'RESPONDER', 'PHP_SELF' => '/manager/index.php', 'REQUEST_TIME_FLOAT' => 1658544812.546169, 'REQUEST_TIME' => [1658544812](tel:1658544812), )
Edit— will update if I make any progress with actual ModX index.php files
I’ve dumped
ini_set('display_errors', '1');
ini_set('display_startup_errors', '1');
error_reporting(E_ALL);
into index.php, and now I see this when I access domain.example.com/manager
Warning: Trying to access array offset on value of type bool in /var/www/domain.example.com/core/vendor/xpdo/xpdo/src/xPDO/Om/xPDOObject.php on line 251
Warning: Trying to access array offset on value of type bool in /var/www/domain.example.com/core/vendor/xpdo/xpdo/src/xPDO/Om/xPDOObject.php on line 251
Should I open an issue at https://github.com/modxcms/xpdo ?
Here is the line of the error
It’s trying to index errorInfo, which is a Boolean.
possibly related: https://github.com/modxcms/xpdo/issues/137
Sure, a summarized report would be good there.
But even with that tracked down, that doesn't explain your empty manager. Does it work better if you disable the error_reporting and display_errors?
If I disable error_reporting and display_errors, then the manager is still empty as before, even after running the setup again with the new mysql user.
oh, I just realized I assume I can do this:
rm -rf /var/www/domain.example.com
unzip modx-3.0.1-pl-sdk.zip
mv modx-3.0.1-pl-sdk /var/www/domain.example.com
cd /var/www/domain.example.com/setup
cp ~/config.xml.backup config.xml
then I drop the MySQL database, create a new empty database called modx_modx, then run the setup command again.
is attempting to re-install modx this way possibly causing me problems, assuming I made no mistakes along the way?
next I will try running this helper script to install modx instead of the default cli instructions.
https://github.com/craftsmancoding/modx_utils/blob/master/installmodx.php
edit: Here are the install logs:
[2022-07-24 11:39:40] (ERROR in xPDO\xPDOConnection::connect @ /var/www/domain.example.com/core/vendor/xpdo/xpdo/src/xPDO/xPDOConnection.php : 89) SQLSTATE[HY000] [2002] No such file or directory
[2022-07-24 11:39:40] (ERROR in xPDO\xPDOConnection::connect @ /var/www/domain.example.com/core/vendor/xpdo/xpdo/src/xPDO/xPDOConnection.php : 89) SQLSTATE[HY000] [2002] No such file or directory
[2022-07-24 11:39:40] (ERROR in xPDO\xPDO::getValue @ /var/www/domain.example.com/core/vendor/xpdo/xpdo/src/xPDO/xPDO.php : 1111) No valid PDOStatement provided to getValue
[2022-07-24 11:39:40] (ERROR in xPDO\xPDOConnection::connect @ /var/www/domain.example.com/core/vendor/xpdo/xpdo/src/xPDO/xPDOConnection.php : 89) SQLSTATE[HY000] [2002] No such file or directory
[2022-07-24 11:39:40] (ERROR @ /var/www/domain.example.com/setup/includes/request/modinstallclirequest.class.php : 298) Could not connect to database server: SQLSTATE[HY000] [2002] No such file or directory
Oh my, I've solved this issue!
I searched for the Connection failed: SQLSTATE[HY000] [2002] No such file or directory error message, and the owner of "Infinity Free" replied that I should replace "localhost" with 127.0.0.1--- they were so certain that they said:
This is 100% guaranteed the one and only case that can cause this error.
I'll make a pull request against this repository, suggesting this replacement:
<database_server>localhost</database_server>
to
<database_server>127.0.0.1</database_server>
This should affect config.dist.new.xml and config.dist.upgrade-advanced.xml.