RAC and network failures

  • From: Ray Stell <stellr@xxxxxxxxxx>
  • To: oracle-l@xxxxxxxxxxxxx
  • Date: Thu, 12 Aug 2004 16:37:00 -0400

10g RAC on Red Hat AS 3.0 using CRS (Cluster Ready Services)

I've was under the impression that RAC was supposed to be able
to handle network failures.  Now that I go back and look, I don't
see this discussed in detail within the documentation.  Rather,
there are hints, such as:


"Through introduction of quorum disk, network failure and node failure
are detected and resolved faster, resulting in faster completion of
cluster reconfiguration."

My testing shows that if a network failure takes place to one node, all
nodes are frozen and basically the concept of high availability is not
on the radar; no new connetions to any node and current connections
to alternate nodes are locked up and end with a 3113 error.

Have others had success in making RAC highly available during
network failures?  Maybe this is a function of not using an OS based
cluster solution.  "Anyone know what this is? Class? Anyone? Anyone? 
anyone seen this before." 

An almost complete tar indicates that this is a bug that would be
treated as a new feature request and NOT something that would happen in
10g if ever.

Ferris Bueller
Ray Stell   stellr@xxxxxx     (540) 231-4109     KE4TJC    28^D
Please see the official ORACLE-L FAQ: http://www.orafaq.com
To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html

Other related posts: