In order to ensure the ability to roll back to our known good environment, I've been asked to set up somewhat complicated sync between three PG nodes. I'll call them CURRENT, TARGET, and FAILSAFE.
CURRENT is the current Prod server, in our datacenter. TARGET is where we intend to migrate to, in Amazon RDS. FAILSAFE is in case we need to roll back to our datacenter. The intended end-result is... CURRENT --syncs with--> TARGET --syncs with--> FAILSAFE ...where each arrow represents a one-way sync with Bucardo. The intention is to cut to TARGET in RDS, but still have the FAILSAFE server in a known good environment in case we need to roll back from RDS, with as close to real-time data as possible. We intend to keep the FAILSAFE server online and synced with TARGET for a few days (weeks?) after the migration to RDS, or until we're confident that RDS performance is sufficient. The reason I haven't simply set up a two-way sync between CURRENT and TARGET is because we don't want the risk of someone making changes on TARGET during testing that would get synced back to CURRENT that could cause problems in live Production. All that backstory is to ask this question: Given that I already have a sync from CURRENT to TARGET, can I safely set up a sync from TARGET to FAILSAFE without disrupting with the first sync? The desired end result would ensure that all three servers are kept in sync until the cutover to TARGET, at which time the sync from CURRENT to TARGET would be disabled, and the sync from TARGET to FAILSAFE would remain in place.
_______________________________________________ Bucardo-general mailing list [email protected] https://mail.endcrypt.com/mailman/listinfo/bucardo-general
