Basically, I am trying to create a DNS cluster using ISPConfig.
There will be two servers, and I would like to be able to update the ISPConfig panel on one server, and have the records duplicated on another, essentially having an ns1 and ns2 setup, administered via one webpage on the primary server.
I currently had two working individually, and so far my efforts have only created a half working scenario. Re-installation of ISPConfig is not a problem if that is what is required.
Any ideas guys?
You dont need a cluster for this. DNS record replication is a builtin feature of ISPConfig and Bind. If you have 2 ISPConfig servers, create a primary dns record on the first server. Then create a secondary dns record on the second server, the details of the dns records will be replicated to the secondary server automatically in future if you change the primary record on the first server.
Thanks for that till.
So I just add the 2nd server in as a secondary into the ISPConfig web, and create a slave on the second, or is it a full record?
The problem I was having, was specifying a primary and secondary via the portal, which propagated the record to the second, but would not update the record on the second when changes were made.
Am I able to do this via a database replication to the other server instead, as I would prefer to only have one site to log into (others will be doing this, and am trying to make it as simple as possible). Such as cron job to the other server?
On the second server you add just a slave record. It gest synced with the primary record on the other server automatically.
Heartbeat & ISP Config Suggestions?
Has anyone got a suggestion on setting up ISPConfig on two servers and using Heartbeat (shared IP) for redundancy and to ease maintenance (using heartbeat an admin can control which server is the 'active' server (serving http/smtp/pop on a virtual IP address much like VRRP or Cisco's HSRP)?
A shared or distributed file system is required or use of a file synchronisation tool (rsync?) to ensure files exist on both servers all the time (excluding temporary process files)? Obviously there is some risk in doing this as temporary files at certain stages of processing (I think maybe inbound SMTP messages doing through virie scanning) may cause the odd problem.
I guess having both servers running SMTP all the time with delivery to a 'shared' location (eg SAN) but just run POP/IMAP/HTTP on the virtual IP address and which ever is the active 'node/server' the service is in a running state whereas on the standby/inactive node/server, the service (eg POP3/IMAP) is not running. Also the MySQL needs to be taken into consideration and I am aware of how MySQL replication (master/slave database servers function).
The goal is either server can completely die (eg, HDD fail) and the other just picks up where the failed server dropped off. I think MySQL replication can be problematic at times when the replication breaks.
I'd appreciate suggestions, what works for you, links to any guides / howto's available, etc?
|All times are GMT +2. The time now is 19:48.|
Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2014, vBulletin Solutions, Inc.