Comments on: First thoughts of Sun Cluster 3.2

By: jodi

jodi — Mon, 14 Jun 2010 06:46:33 +0000

Hii, anyone can help me regarding. how to configure oracle database as scalable data service ?

By: Dr. Kenneth Noisewater

Dr. Kenneth Noisewater — Mon, 11 May 2009 17:10:00 +0000

BTW, just took the 3.2 class, very neat stuff.. However I was wondering if anyone had any hints of a timeframe for if/when ZFS will be globalized? Or are they thinking of making it clusterable (say a clustered zpool on shared storage that arbitrates at the zpool level, presenting zfs filesystems that are global)?

‘cuz that is probably the one thing that keeps VCS from getting the boot in my org: CFS…

By: Tirthankar

Tirthankar — Sun, 01 Feb 2009 08:28:30 +0000

Solaris Cluster 3.2 U2 has been released and comes with many more useful features like providing the ability to create zone clusters and running ORACLE RAC inside a zone.

Refer to the video blog about SC3.2U2
http://blogs.sun.com/SC/en_US/entry/sc32u2_now_available

Also Refer to SC success story at EMBARQ
http://www.sun.com/customers/software/embarq.xml

By: JD

JD — Wed, 02 Jul 2008 01:04:14 +0000

Any one have bad VCS experience on Non-Sun systems? currently evaluating solutions

By: Konstantin

Konstantin — Tue, 05 Feb 2008 21:12:33 +0000

To Joe:
/globaldevices has to be on EVERY cluster node. and yes, its in the docs.

To Paul:
The Resource (sun cluster term) is SharedAddress.

- Konstantin

By: Paul

Paul — Mon, 28 Jan 2008 22:30:52 +0000

Hi Matty:

Great blog! Any idea how to get a 2-node Sun Cluster to provide a round-robin Virtual IP address similar to what a load balancer would do? (i.e. when both nodes are running fine, load is balanced but when one is down all traffic goes to the other?)

Thanks!

Paul

By: Joe

Joe — Wed, 23 Jan 2008 23:53:35 +0000

Matty, Like the Blog,
Got a question, does the GlobalDevices slice have to be on each of the local nodes or can it be on Global Shared Storage like the quorum slice in a two node config?
I’m just now starting to install cluster 3.2 for the first time and this question doesn’t really get answered in the docs.

-Joe

By: matty

matty — Thu, 03 May 2007 01:46:26 +0000

Hi Dean,

A centralized syslog solution would work for the cluster engine logs, but coulnd’t be used for the agent and command logs. In regards to point number 2, the sczsmf and sczsh resources can greatly assist with this. I am currently working on a project that is deploying highly available zones, and we are looking to use the Sun cluster sczsh or sczsmf components to start, stop and monitor the applications running inside these zones.

- Ryan

By: dean ross-smith

dean ross-smith — Mon, 30 Apr 2007 14:01:56 +0000

as far as logging goes, how ’bout setting up a centralized logging server through syslog.conf? I get all of my cluster messages on a host that’s external to the cluster in one /var/adm/messages file …

A couple of things that seem to be cool for creating failover type services… I want to try this but haven’t yet… going off of what I’ve read…
1. any service that’s setup through the service framework (uses xml for controlling start and stop of a daemon instead of the rc and init.d directories) can be setup as a failover service- no cluster agent necesssary. That’s cool.
2. Because zones are supported as failover, that means that we should be able to set up any non-cluster aware application with init scripts for starting and killing it and fail the zone around a cluster with non-cluster aware applications inside. That seems useful to me…