Find the answer to your Linux question:
Results 1 to 4 of 4
Today we have an expensive solution for our CMS/E-commerce site. We use Proprietary software and solutions, and for that we pay a hefty license fee for several products, tools and ...
Enjoy an ad free experience by logging in. Not a member yet? Register.
  1. #1
    Just Joined!
    Join Date
    Aug 2006
    Posts
    7

    No singel point of failure and failsafeclusters


    Today we have an expensive solution for our CMS/E-commerce site.
    We use Proprietary software and solutions, and for that we pay a hefty license fee for several products, tools and databases. Even our hosting is more expensive due to the nature of license fee.

    We are actually currently very pleased with the level of professionalism we recieve from our host, and the response time is good. Our only problem is the price, and the fact that we inhouse also use Proprietary software.

    As a counter messure we have now started looking for open source solutions mainly GPL and BSD tools and software. I am sure as of today we will find that the linux world will meet every demand we have.

    But for the hosting part, we would like to set up something like this:

    This will be achieved by using 3 dedicated webservers for files etc, and 2 or 3 dedicated servers for the My SQL. I know the pic just shows 2 webservers and two DB servers but that is merly for illustrating my point.

    Our CMS/e-commerce site have high traffic, and for that part it is importent that we have responsive system, with no lag even with high numbers of simultanous users are logged in. Also we will try to focus on a failsafe(correctname?) cluster that uses loadbalancing.

    We will happily recieve tips on what software to use for loadbalancing
    and also for clustering. Preferrable free, or low cost solutions. Allthough other alternatives might be considered as well if they have
    huge advantages over the free ones. Also tips on how to set up the nodes so that if one node goes down, other will take over til problem is fixed.

    Als I wonder if it is possible to gain any of the advantages of virtualization software such as VM ware of virtuzzo with a singel solution spread over several servers. Primarly I am thinking of the possibility to easily have one node up and running if system errors occours.

    • Multiple connections to more then one ISP

    • No singel point of failure, means redundant disks, powerunits etc

    • Loadbalancing to assure resposiveness, fastloading pages and best nhandling of userrequests

    • 2 or 3 servers for webpages

    • 2 or 3 servers for My SQL server


    So I am looking for bestpractices for:
    1. Loadbalancing
    2. Clustering
    3. Replace any failed nodes automatically
    4. Make sure there is enuogh cpu available to remain responsive
    5. Recive tips on how to achive these goals
    6. Make sure the DB is up for lookups, processing, and queries
    7. Is it possible to use virtualization software in a case like this, and will it give us anay benefits?

  2. #2
    Linux Enthusiast
    Join Date
    Dec 2004
    Posts
    637
    VMware EX handles this especially with their vmotion technology. We use this and it is awesome.

  3. #3
    Just Joined!
    Join Date
    Aug 2006
    Posts
    7
    Thx for the reply.

    I belive the overhead for the Vmware Ex is less then 2%? Can you tell me if the Ex uses lots of rescources? The Ex is lower level then the OS if memory servers me correct?

    Did you read any cases before you implemented this? If so could you post a link to them? I am looking for success stories, whitepapers etc.

  4. $spacer_open
    $spacer_close
  5. #4
    Linux Enthusiast
    Join Date
    Dec 2004
    Posts
    637
    The best place is to read the various white papers that are available at http://www.vmware.com The largest deployment that I'm aware of is Nationwide Insurance.

    As for ESX, it is a more down to the bare metal deployment as compared to running atop a conventional OS. ESX can hot failover and re-provision failed server instances in real time without loss of availability.

    As for resource utilization, well, it depends upon your guest configuration. Our standard server deployment has 24 GB of RAM and with this, we run no more than six guests - allowing for greater than 2 GB RAM per virtual OS.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •