• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 236
  • Last Modified:

how to make clustering in linux

Hi all,
my manager assign me the task of making a cluster on a two lans one of them use centos and the other use fedora and they are separte but each one of them working in a  tow modes one of them is a mixed network with windows and the other is a pure linux network.
the problem is I'm a newbie in that task I have no experience in that area before so can any body help me in that???
any help, idea, note, hint, examples, turorials, links or books are very welcomed
please reply as soon as possible (I need your help urgently)
thanks in advance
best Regards
  • 2
1 Solution
Ok, the first question is... what do you want to cluster? Storage? Services? Very difficult to point you in the right direction as the angle from which you attack this will vary significantly depending on what you're trying to do.

One thing I would say though, is it's not compulsary, but recommended to use the same version/distro of Linux on all cluster nodes. That way, if you're using software like Heartbeat, you'll have the same build of the same version etc..
I agree. Without understanding of the thing to cluster, you cannot HAVE a cluster. On this later. I would like to stress the need for similarity between the cluster members (nodes). You don't want to go into a cluster with different node types.

About clusters:
A cluster is not just a "throw this software on these few servers and let's get some beer", but it is a game of risks and assumptions. You must setup an SLA you expect to commit - What is the maximal offline time you can allow yourself to accept? What is your expected expenses when assuming that cluster multiplies most of your existing hardware. For example - if you have a cluster between two nodes, but both are connected to each other, and to the rest of the network by a single switch, you have a single point of failure - the switch. Turn it off (or burn its power supply), and you have zero availability.
Same goes for storage devices, for systems - having two low-budget low-availability servers is, on most cases, worse than having a single reliable server. Same goes for connection to the storage - Use a single cable, or a single fiber switch, and you get a single point of failure. Yada yada yada. You get the point. Cluster without a risk management (much like backups) has no value, and would serve no high availability when the need will rise.
shang3000Author Commented:
hi alextoft,
hi ezaton,
thanks for reply
I need cluster for a web portal using apache, mysql I'm newbie so I need your suggestion to make cluster that is reliable and have load balance and your recommendations (what distribution to use? does it's a good decision to have fedora as this cluster distribution?)
please help me in that
best regards
I would use something supported, like RHEL, Suse, or the likes, and nothing of the sort of Fedora.
This is a load-balancing cluster, assuming all data is saved only in the database. You could, otherwise, use one node as a MySQL server, and two nodes as web servers, with shared storage (concurrent access, like GFS or OCFS2) for the web content.

Either case, this require more planning than can be supplied in this forum.

Featured Post

Concerto Cloud for Software Providers & ISVs

Can Concerto Cloud Services help you focus on evolving your application offerings, while delivering the best cloud experience to your customers? From DevOps to revenue models and customer support, the answer is yes!

Learn how Concerto can help you.

  • 2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now