Content Failover with Primary and Secondary Clusters
This discussion thread has been imported from the Centera Developer Portal Forums.
The first message is shown here, and the complete thread is in the attached PDF file.
Please read the PDF and continue the discussion here.
From jjdemski Posted: Sat Sep 20, 2008
I'm testing content failover in a configuration with a primary and a secondary cluster. In one case, the clip in question exists only on the secondary cluster. Content failover works for read, exists, and delete: they all see the clip on the secondary cluster.
However, if the primary cluster is running CentraStar 3.1.3, content failover works on read, but it takes almost a minute for the read to complete. (Exists and delete are fast.) If the primary cluster is running CentraStar 4.0.1, read is fast.
I think I noticed another slow content failover in my testing, but I was more interested in functionality than performance, and didn't make note of it.
Are there any recommendations of what versions of CentraStar should be used on a primary cluster to avoid these slow content failovers? Is there a list of what does or doesn't work well?