e i g h t - c u b e d . c o m 

Continuous Access misconfiguration

This company decided to enter into an agreement with a third party disaster recovery supplier to supply storage, rack space, and a managed solution. At my suggestion (made back in 2009 when I was performing the IA64 port for them) the OpenVMS solution uses Continuous Access.

Once this got going, I was unfortunately not consulted about further OpenVMS requirements, and so was had some interesting issues the first time we attempted a disaster recovery test with this setup.

Recently, the Windows guys have been attempting to replicate some large volumes with CA, and have been unable to get the replication group to be managed by the "B" controller.

It turns out that when the third party configured CA, they seriously misconfigured the host port priority screen. Thanks to some support from a mate in the States, I've sorted them out.

The best thing you can do on the "Host port data replication settings" screen is to hit the "Reset to defaults" button, and let the controllers figure it out (or if you know your exact wiring scheme, you can specify priorities).

In our case, we have only a single port wired on each local controller dedicated to CA. The far end is wired redundantly. So this makes it easy, with preferences being local controller A to remote controller A as the first priority, and A to B as the second, then local controller B to remote controller B as the first, and B to A as the second.