Clustering And Ha Planning
Hello , CentOS users:
My name is Leandro, I have been using CentOS for 4 years and this is the first post in this mail list. I would like to study and introduce myself in clustering and high availability for CentOS, currently I have not experience at all about it. I would like to ask about the newest method to achieve high availability
, load balancing on linux / CentOS. So far I have seen the Clustering docs writen for CentOS 5 and the HA
documentation from www.linux-ha.org that have been published in 2010. So, I would like to ask to comunity, which are the new methods for clustering and get HA and where to get updated documentation.
Regards, Leandro.
4 thoughts on - Clustering And Ha Planning
Hello Leandro,
CentOS 5 is quite old and different from current CentOS 7, some things have changed, mostly improved and as usual your favourite search engine is your friend. e.g. http://clusterlabs.org/quickstart-redhat.html https://skcave.wordpress.com/2014/11/04/creating-high-availability-cluster-with-CentOS-7/
I contend the appropriate approach to HA should be based on what services you need to keep available. an HA file server has quite different requirements and implementations than a HA relational database server or a HA DNS server. There’s no magic one size fits all solutions.
Thanks for pointing that. I would like to learn about clustering and HA, so if I have to chose a service for my testing scenario It will be a radius or a mysql justo to keep it simple.
Leandro.
The main mailing list for HA clustering in “Clusterlabs Users”:
http://clusterlabs.org/mailman/listinfo/users.
It’s not strictly for any OS, but RHEL/CentOS and SUSE are probably the most common OSes.
I might recommend starting with this:
https://alteeve.ca/w/History_of_HA_Clustering
The Linux-HA project (heartbeat) is long deprecated. The stack to learn is Corosync + Pacemaker. As Nux mentioned, CentOS 7 is the best platform. As you’ll see in the History link above, there was a lot of changes that happened between 2008 ~ 2013 era. Learning on any older CentOS means you’re learning an old stack, which probably doesn’t make sense outside of a few cases.
We also have an active IRC channel on #clusterlabs on freenode.net, too. If you stop by, be sure to idle. Folks are from all over so different people are around at different times. That said, people are good about replying to questions when they come around.
Welcome to HA!
digimer