Our plan for deployment is exactly as Clark described: * Several API nodes behind a load balancer * PostgreSQL master/slave replication * HSMs in HA paired mode * Several Worker nodes I’m also curios as to why this would be considered “clunky”? -Doug On 3/19/14, 1:21 PM, "Clint Byrum" <clint at fewbar.com> wrote: >Excerpts from Clark, Robert Graham's message of 2014-03-19 07:41:35 -0700: >> Has there been much discussion on how to ensure that keys are >> recoverable in the event that Barbican has some sort of horrific >> failure? >> >> I suppose a HA frontend, Redundant Keystore Databases and HA paired HSMs >> would be the most obvious non-code-writing path but this feels pretty >> clunky, I was wondering if it had been discussed yet? Possibly it should >> be something for a design session? >> > >Sorry, what is clunky about backing up your data? > >_______________________________________________ >Mailing list: >http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack >Post to : openstack at lists.openstack.org >Unsubscribe : >http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 5660 bytes Desc: not available URL: <http://lists.openstack.org/pipermail/openstack/attachments/20140319/d33b5865/attachment.bin>