Estou tentando executar o hub Mercure do supervisor, mas não é possível para mim. O Mercure está na mesma máquina do servidor web com o host virtual SSL para pami54.local
domínio.
[program:mercure]
environment=JWT_KEY="m3rcu353cr37pa55pra53DEV"; CORS_ALLOWED_ORIGINS="https://pami54.local"; PUBLISH_ALLOWED_ORIGINS="*"; ADDR="pami54.local:3000"
command=/home/frizquierdo/mercureLinux/mercure run -config /home/frizquierdo/mercureLinux/Caddyfile
process_name=%(program_name)s_$(process_num)s
numprocs=1
autostart=true
#directory=/tmp
autorestart=true
startsecs=5
startretries=10
redirect_stderr=false
stdout_capture_maxbytes=1MB
stderr_capture_maxbytes=1MB
stdout_logfile=/var/log/supervisor/mercureout.log
stderr_logfile=/var/log/supervisor/mercureerror.log
EDITAR:
Eu fundei uma solução parcial. No Caddyfile desativei o servidor caddy na porta 80, definindo a diretiva auto_https como 'disable_redirects', ficando o Caddyfile desta forma:
# Learn how to configure the Mercure.rocks Hub on https://mercure.rocks/docs/hub/config
{
{$GLOBAL_OPTIONS}
auto_https disable_redirects
}
pami54.local:3000
log
tls /etc/apache2/ssl-cert/pami54.local.crt /etc/apache2/ssl-cert/pami54.local.key
route {
encode zstd gzip
mercure {
# Transport to use (default to Bolt)
transport_url {$MERCURE_TRANSPORT_URL:bolt://mercure.db}
# Publisher JWT key
publisher_jwt {env.MERCURE_PUBLISHER_JWT_KEY} {env.MERCURE_PUBLISHER_JWT_ALG}
# Subscriber JWT key
subscriber_jwt {env.MERCURE_SUBSCRIBER_JWT_KEY} {env.MERCURE_SUBSCRIBER_JWT_ALG}
# Extra directives
cors_origins https://pami54.local
publish_origins *
{$MERCURE_EXTRA_DIRECTIVES}
}
respond /healthz 200
respond "Not Found" 404
}
Sua configuração do virtualhost:
<IfModule mod_ssl.c>
#SSLStaplingCache "shmcb:${SRVROOT}/logs/ssl_stapling(32768)"
<VirtualHost *:443>
ServerName pami54.local
ServerAlias wwww.pami54.local
DocumentRoot "/var/www/html/pami54.local/public"
DirectoryIndex index.php
<Directory "/var/www/html/pami54.local/public/">
AllowOverride All
Order Allow,Deny
Allow from All
#Require local
Require all granted
<IfModule mod_rewrite.c>
Options -MultiViews
RewriteEngine On
RewriteCond %{REQUEST_FILENAME} !-f
RewriteRule ^(.*)$ index.php [QSA,L]
</IfModule>
</Directory>
SSLEngine on
SSLProtocol all -SSLv3 -SSLv2
SSLCipherSuite ALL:!ADH:!EXPORT56:RC4+RSA:+HIGH:+MEDIUM:+LOW:+SSLv2:+EXP:+eNULL
SSLCertificateFile "/etc/apache2/ssl-cert/pami54.local.crt"
SSLCertificateKeyFile "/etc/apache2/ssl-cert/pami54.local.key"
SSLUseStapling off
<FilesMatch "\.(cgi|shtml|pl|asp|php)$">
SSLOptions +StdEnvVars
</FilesMatch>
BrowserMatch ".*MSIE.*" nokeepalive ssl-unclean-shutdown downgrade-1.0 force-response-1.0
SSLProxyEngine On
ProxyRequests Off
ProxyPreserveHost On
#ProxyPass '/.well-known/mercure' 'https://pami54.local:3000/.well-known/mercure' connectiontimeout=300 timeout=300
ProxyPass '/.well-known/mercure' 'https://pami54.local:3000/.well-known/mercure'
ProxyPassReverse '/.well-known/mercure' 'https://pami54.local:3000/.well-known/mercure'
</VirtualHost>
</IfModule>
Desta forma, os clientes podem se conectar ao Mercure on https://pami54.local/.well-known/mercure
, mesmo quando ocorre uma desconexão, o cliente se reconecta com sucesso ao hub, mas quando o servidor (webapp) tenta publicar uma notificação no hub, erro de log do cliente http do symfony:
[2022-04-27T19:29:40.857698-04:00] http_client.INFO: Request: "POST https://pami54.local/.well-known/mercure" [] []
[2022-04-27T19:29:40.871491-04:00] messenger.WARNING: Error thrown while handling message App\Message\NotificacionMarcarComoLeidaMessage. Sending for retry #1 using 1000 ms delay. Error: "Handling "App\Message\NotificacionMarcarComoLeidaMessage" failed: Failed to send an update." {"message":{"App\\Message\\NotificacionMarcarComoLeidaMessage":[]},"class":"App\\Message\\NotificacionMarcarComoLeidaMessage","retryCount":1,"delay":1000,"error":"Handling \"App\\Message\\NotificacionMarcarComoLeidaMessage\" failed: Failed to send an update.","exception":"[object] (Symfony\\Component\\Messenger\\Exception\\HandlerFailedException(code: 0): Handling \"App\\Message\\NotificacionMarcarComoLeidaMessage\" failed: Failed to send an update. at /var/www/html/pami54.local/vendor/symfony/messenger/Middleware/HandleMessageMiddleware.php:129)\n[previous exception] [object] (Symfony\\Component\\Mercure\\Exception\\RuntimeException(code: 0): Failed to send an update. at /var/www/html/pami54.local/vendor/symfony/mercure/src/Hub.php:104)\n[previous exception] [object] (Symfony\\Component\\HttpClient\\Exception\\TransportException(code: 0): The request was not processed and can be safely retried at /var/www/html/pami54.local/vendor/symfony/http-client/Response/CommonResponseTrait.php:148)\n[previous exception] [object] (Symfony\\Component\\HttpClient\\Exception\\TransportException(code: 0): The request was not processed and can be safely retried at /var/www/html/pami54.local/vendor/symfony/http-client/Chunk/ErrorChunk.php:65)\n[previous exception] [object] (Amp\\Http\\Client\\Connection\\UnprocessedRequestException(code: 0): The request was not processed and can be safely retried at /var/www/html/pami54.local/vendor/amphp/http-client/src/Connection/DefaultConnectionFactory.php:117)\n[previous exception] [object] (Amp\\Http\\Client\\SocketException(code: 0): Connection to 'pami54.local:443' failed at /var/www/html/pami54.local/vendor/amphp/http-client/src/Connection/DefaultConnectionFactory.php:118)\n[previous exception] [object] (Amp\\Socket\\ConnectException(code: 111): Connection to tcp://pami54.local:443 refused at /var/www/html/pami54.local/vendor/amphp/socket/src/DnsConnector.php:108)"} []
O que esse erro significa que está lançando o componente HTTP CLIENT do symfony? Achei que tinha resolvido o problema. Pelo menos os clientes conectam e reconectam ao hub, agora o problema é quando o aplicativo da web tenta publicar no hub Mercure.
Devo dizer que meu ambiente local não possui um servidor DNS, tudo está com host virtual local e nome de domínio declarado na /etc/hostname.conf
máquina virtual Debian:
#/etc/hostname.conf
debiandev
pami54.local
Responder1
Eu tenho solução. No Caddyfile desativei o servidor caddy na porta 80 atribuindo qualquer outra porta, por exemplo 8081, e definindo auto_https
a diretiva para 'disable_redirects', permanecendo Caddyfile
assim:
# Learn how to configure the Mercure.rocks Hub on https://mercure.rocks/docs/hub/config
{
{$GLOBAL_OPTIONS}
http_port 8081
auto_https disable_redirects
}
pami54.local:3000
log
tls /etc/apache2/ssl-cert/pami54.local.crt /etc/apache2/ssl-cert/pami54.local.key
route {
encode zstd gzip
mercure {
# Transport to use (default to Bolt)
transport_url {$MERCURE_TRANSPORT_URL:bolt://mercure.db}
# Publisher JWT key
publisher_jwt {env.MERCURE_PUBLISHER_JWT_KEY} {env.MERCURE_PUBLISHER_JWT_ALG}
# Subscriber JWT key
subscriber_jwt {env.MERCURE_SUBSCRIBER_JWT_KEY} {env.MERCURE_SUBSCRIBER_JWT_ALG}
# Extra directives
cors_origins https://pami54.local
publish_origins *
{$MERCURE_EXTRA_DIRECTIVES}
}
respond /healthz 200
respond "Not Found" 404
}
Agora é possível que os clientes se conectem ao hub, mas a aplicação web não consegue publicar nele.
Eu atualizei a postagem.