08-11-2010
HA Cluster solution for Linux - which one to use?
Hi all experts,
I need your kind suggestions or feedback on choice of clustering software for red hat linux nodes running certain applications which I need to make highly available.
Minimum requirement is 2 nodes; all nodes shall be active-active running distinct applications e.g. node 1 runs application ABC and node 2 runs application PQR, node 1 is back up node for PQR and node 2 for ABC. Node 1 and node 2 do not have any shared storage required, they access back end database which is protected by Oracle's dataguard.
I need to have cost-effective HA with minimum downtime. There are various HA solutions available in market - SteelEye, Veritas Cluster, Red Hat cluster suite, IBM HACMP etc. and I have almost narrowed down to Red Hat Cluster(for it being cheaper) vs Veritas (for it being feature rich).
Any experiences with clustering? Any pros and cons of these two cluster software (with references would be appreciated) would be of great help?
Dear mods, if this thread is not suitable for this sub-forum; kindly advise where to post within this forum
5 More Discussions You Might Find Interesting
1. UNIX for Advanced & Expert Users
Dear All Experts,
Would like to know the maturity/ stability of Redhat Linux AS 3.0 and Solaris.
My organization need to setup cluster solution. We are well-versed with Veritas Cluster on Solaris.
We are thinking of waiting for certification support of the various ISV like Oracle, Veritas... (3 Replies)
Discussion started by: izy100
3 Replies
2. Linux
All,
I am most familiar with Solaris, and I am in the process of learning Linux (Fedora 5), and one of my tasks is to replace our current NAS solution. We currently use EMC Celerra, but it is way too expensive for what we use it for. So I have looked into Linux.
We mostly we have a Windows... (1 Reply)
Discussion started by: kjbaumann
1 Replies
3. Red Hat
Hi ,
I am a complete newbie to linux, I have been given a job to install a mail solution (postfix if possible) I would like to know how should I go about installing..)
Can any one provide me some steps to go about it..
i have read about postfix and the installation steps, it seems there are... (1 Reply)
Discussion started by: anilhk
1 Replies
4. High Performance Computing
Hi, I am not sure if the term "cluster" fits my situation or not. I have 5 Linux computers, each has different a host-name, and my users have to remember those the computer names to login and find out which computers have free CPUs, so they can run their jobs on those computers have free CPUs.
... (4 Replies)
Discussion started by: hiepng
4 Replies
5. Red Hat
Can some one tell me free cluster solution avaialable for RHEL5 we just want to test the clustering on RHEL. (6 Replies)
Discussion started by: fugitive
6 Replies
LEARN ABOUT DEBIAN
clmresidue
clm residue(1) USER COMMANDS clm residue(1)
NAME
clm residue - extend clustering of subgraph to clustering of graph.
clmresidue is not in actual fact a program. This manual page documents the behaviour and options of the clm program when invoked in mode
residue. The options -h, --apropos, --version, -set, --nop are accessible in all clm modes. They are described in the clm manual page.
SYNOPSIS
clm residue -icl fname (input clustering) -imx fname (input graph) [-rpm fname (residue projection matrix)] [-o fname (output clustering
file)]
DESCRIPTION
The cluster file presumably contains a clustering of a subgraph of the graph G contained by the matrix file, implying that the row domain of
that clustering is a subset of the node domain of the graph. clm residue will compute a simple upwards projection of that clustering,
resulting in a clustering of G.
OPTIONS
-imx fname (input graph)
-icl fname (input clustering)
-rpm fname (residue projection matrix)
The name for the file in which the residue projection matrix is stored. It contains for each node and for each cluster in the input clus-
tering the combined weight of the (edges to the) neighbours of that node in that cluster. The default output file name is out.rpm.
-o fname (output clustering file)
The name for the file in which the majority vote projection clustering is stored. Aka the majority vote projection clustering storage file
name.
This embodies a very crude approach to distributing the missing nodes in the input clustering onto that same cluster. For each node, the
cluster is taken for which the residue projection is the largest (see above). The residue (i.e. the set of missing nodes) is initially
viewed as a cluster; the set of nodes for which the projection is largest for the residue itself will form a new cluster, if it is not
empty.
By default output is sent to STDOUT.
AUTHOR
Stijn van Dongen.
SEE ALSO
mclfamily(7) for an overview of all the documentation and the utilities in the mcl family.
clm residue 12-068 8 Mar 2012 clm residue(1)