Ok, I have logical replication up-and-running (I guess - seemed to simple to be working. Shouldn’t it be complicated, requiring many steps and configuration changes?), maxing out one CPU core on each machine (more or less), and showing network throughput of around 15M. If DU changes are to be believed, it’s transferring data at about 0.8GB/minute, implying something like a 8 hour transfer time. Of course, since it is replication, it has the benefit that any data that comes in during that 8 hour window should also be replicated, after which the two systems should remain in sync allowing for zero (or nearly so) downtime cutover. Which is nice. Any gotchas I need to be aware of during this initial transfer window, such as WAL files building up on the source machine? --- Israel Brewster Software Engineer Alaska Volcano Observatory Geophysical Institute - UAF 2156 Koyukuk Drive Fairbanks AK 99775-7320 Work: 907-474-5172 cell: 907-328-9145
|