topic Re: clustering in Operating System - Linux

clustering

gauravsingh235 — Thu, 05 Apr 2007 02:31:16 GMT

How to configure a fail over / HA cluster with postgresql in LINUX.

Basic ideas of what is Quorom,base volumes in SANBOX

Re: clustering

Steven E. Protter — Thu, 05 Apr 2007 02:38:59 GMT

Shalom,

What kind of Linux.

If you use redhat, you can use RHCS, Red Hat Clustering, which has a new qdisk(quorums feature) or you can use HP Serviceguard for Linux.

Serviceguard is pretty stable and uses the quorum disk concept.

SEP

Re: clustering

Serviceguard for Linux — Fri, 06 Apr 2007 09:27:55 GMT

Serviceguard actually has multiple quorum options. For two node clusters one is disk based and the other is network based (application runs on some third system in your environment).

Your choice :)

Re: clustering

gauravsingh235 — Mon, 09 Apr 2007 03:26:01 GMT

My situiation is some thing Like this

There will be some files continuously (24*7)
uploaded to the storage box and Uploaded Files info are loaded into the Database server(Postgresql) so that i can query the infos from database server and make it in a viewable format for my WEB interface.

We are planning to put db1 and db2 as two database server in cluster with mirroring.
If i am planning for mirroring cluster will it affect my processing? will it take much time to update database continuously, because files are comming continuosly (24*7)

Help me or suggest a solution for this

Re: clustering

Ivan Ferreira — Mon, 09 Apr 2007 08:29:44 GMT

Probably, your best option is to use a shared storage with a cluster file system, like Red Hat GFS (CentOS GFS), Oracle CFS2 is also posix compliant.

Another option is to use a third server (or servers) to act as NFS servers for the database servers, and files are uploaded to the NFS servers. You can also use GNDB devices.