The Community Forums

Interact with an entire community of cPanel & WHM users!
  1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

fault tolerance/redundancy/clustering

Discussion in 'Bind / DNS / Nameserver Issues' started by tdens, Aug 19, 2011.

  1. tdens

    tdens Member

    Joined:
    Aug 18, 2011
    Messages:
    19
    Likes Received:
    0
    Trophy Points:
    1
    Hi Folks,

    How do most people provide redundancy and/or load balancing to their customers when using WHM/cpanel?

    I'm looking at setting WHM/cpanel up for our web customers, but I'd like some laod balancing and/or at least fault tolerance. How I usually handle this is to set things up in either an active/passive config using heartbeat/corosync and drbd, or setup multiple VM instances of the same server up behind a load balancer using remote mysql and NFS to allow all VMs to access the same data. What partitions on the server would need to be shared for either of these scenarios to be feasible? So far I have:

    /home
    /var/cpanel

    There are also several databases I'd need to migrate to the remote mysql server, and a few log files in /var/lib/mysql. Cphulk and leechprotect were easy enough to move, but there's quite a few services that simply report that mysql is off, and I'd have to turn it on locally to view whatever it is that I want to see. On a side note, I find it strange that these wouldn't automatically be moved to the remote server. I find it even stranger that I'm not given a screaming red popup when I setup a remote sql server that I'm disabling WHM's built in cphulk protection.

    I can only assume what I'm trying to do is not the standard way things are done - so maybe the real question is my first one - how do people generally offer fault tolerance and load balancing to customers when using WHM/cpanel?

    Any help with either question, even pointers to docs, would be appreciated - THANKS!
     
  2. tdens

    tdens Member

    Joined:
    Aug 18, 2011
    Messages:
    19
    Likes Received:
    0
    Trophy Points:
    1
    Apparently this variety of HA isn't supported or recommended. Having had the pleasure of dealing with mail servers for several years, I can't see telling customer that their mail server is going to be down for the next several hours and that they've probably lost half a day's mail (not to mention dealing with the nightmare of out-of-sync UIDL files) going very well.

    Do people really stand up a single server or VM that houses customer sites and *mail* (and possibly DNS and sql if those haven't been spun off) and just cross their fingers that boxen don't die and that proc/RAM/network don't flatline up-and-to-the-right? Obviously, housing your vmdks on a SAN can mitigate some of this, but not the load balancing portion. Is there a good rule of thumb for domains-per-instance of WHM I can fall back to?
     
Loading...

Share This Page