Multiserver setup, users not be replicated to other servers

Discussion in 'General' started by webguyz, Jul 19, 2014.

  1. webguyz

    webguyz Member HowtoForge Supporter

    I added a new ftp user and when I tried to use it kept getting a failed password. Finally checked the web server that I had added the new ftp user to and it was in the master dbispconfig db , but not in the remote servers dbispconfig db and that's why it was failing. I had recently updated my setup with the latest patch, but had not tried adding anything new since then.

    What forces the dbispconfig db's to update from the master?

    Thanks!
     
  2. till

    till Super Moderator Staff Member ISPConfig Developer

    The slaves are pulling the updates from the master. either the server.sh cronjob on the slave is not running or the slave is not able to connect the master mysql server anymore to pull the updates.
     
  3. webguyz

    webguyz Member HowtoForge Supporter

    I ran the server.sh manually and got this error. Any idea what might have happened?

    /sbin/iptables
    /sbin/ip6tables
    19.07.2014-15:35 - ERROR - Replication failed. Error: (web_domain) in MySQL server: (localhost) Unknown column 'backup_excludes' in 'field list' # SQL: REPLACE INTO web_domain (`domain_id`,`sys_userid`,`sys_groupid`,`sys_perm_user`,`sys_perm_group`,`sys_perm_other`,`server_id`,`ip_address`,`ipv6_address`,`domain`,`type`,`parent_domain_id`,`vhost_type`,`document_root`,`web_folder`,`system_user`,`system_group`,`hd_quota`,`traffic_quota`,`cgi`,`ssi`,`suexec`,`errordocs`,`is_subdomainwww`,`subdomain`,`php`,`ruby`,`python`,`perl`,`redirect_type`,`redirect_path`,`seo_redirect`,`ssl`,`ssl_state`,`ssl_locality`,`ssl_organisation`,`ssl_organisation_unit`,`ssl_country`,`ssl_domain`,`ssl_request`,`ssl_cert`,`ssl_bundle`,`ssl_key`,`ssl_action`,`stats_password`,`stats_type`,`allow_override`,`apache_directives`,`nginx_directives`,`php_fpm_use_socket`,`pm`,`pm_max_children`,`pm_start_servers`,`pm_min_spare_servers`,`pm_max_spare_servers`,`pm_process_idle_timeout`,`pm_max_requests`,`php_open_basedir`,`custom_php_ini`,`backup_interval`,`backup_copies`,`backup_excludes`,`active`,`traffic_quota_lock`,`fastcgi_php_version`,`proxy_directives`,`last_quota_notification`,`rewrite_rules`,`added_date`,`added_by`) VALUES ('29','23','23','riud','ru','','2','*','','rbrmx.net','vhost','0','name','/var/www/clients/client22/web29','','web29','client22','5000','-1','n','n','y','0','1','www','fast-cgi','n','n','n','','','','n','','','','','','','','','','','','','webalizer','All','','','n','dynamic','10','2','1','5','10','0','/var/www/clients/client22/web29/web:/var/www/clients/client22/web29/private:/var/www/clients/client22/web29/tmp:/var/www/rbrmx.net/web:/srv/www/rbrmx.net/web:/usr/share/php5:/usr/share/php:/tmp:/usr/share/phpmyadmin:/etc/phpmyadmin:/var/lib/phpmyadmin','','none','1','','y','n','','','','','0000-00-00','')
    19.07.2014-15:35 - ERROR - Error in Replication, changes were not processed.
    finished.
     
  4. webguyz

    webguyz Member HowtoForge Supporter

    I checked the dbispconfig DB on the master and it does have a backup_excludes field in the web_domain table but none of my slave servers have that field and that's why my replication is failing. Did something change in the last update that added that field? Do I need to go and manually add that field into web_domain table manually on each slaves dbispconfig file?

    Thanks!
     
  5. till

    till Super Moderator Staff Member ISPConfig Developer

    The database layout has changed in the last update, but there are no manual changes required, you just have to update always all nodes of a cluster so that they run the same ispconfig version. The master should be updated first and afterwards each slave has to be updated. On one of the salves you should choose "reconfigure permissions in master database", it does not matter on which of the slaves.
     
  6. webguyz

    webguyz Member HowtoForge Supporter

    I missed that point. For some reason I thought you only had to run the update on the master only. So if I run the update on the slaves then everything should be back working again.

    Thanks for your excellent support Till!
     

Share This Page