On Thu, Feb 7, 2019 at 3:09 AM David Christensen <[email protected]> wrote:
> > We're moving our EC2 Postgres 1TB DB to RDS and decided to use Bucardo, > as it supports multiple PG versions. > > We have a PG 9.2 cluster, and instead of upgrading it, we'll use Bucardo > to replicate to a PG 9.6, then use DMS or pg_dump to restore it in RDS. > > > > Because of the size of the current 9.2 DB, I cannot stop the application > so it doesn't write to the DB while I do pg_dump and restore it in the new > bucardo slave. > > > > So, I thought I would do something like this: > > • Install Bucardo and add large tables to a pushdelta sync > > • Copy the tables to the new server (e.g. with pg_dump) > > • Startup Bucardo and catch things up (e.g. copy all rows changes > since step 2) > > Steps for the above would be pretty much like this article; if I'm not > mistaken. > > > > My questions are: > > • Is it the right approach? do you guys have any other suggestions? > > Yeah, as long as you’re doing a one-way sync (master -> target) it’s > sufficient to start capturing the deltas (generally by creating a sync with > auto-kick off), then pg_dump the data, then kick the sync/set autokick > until you’re caught up then cutover the app as needed. > > First of all, thanks for your reply! That would be pushdelta <https://bucardo.org/pushdelta/>, correct? > > • I'll have the following: pg-9.2 master --> bucardo instance > (with the bucardo DB only) --> pg-9.6 slave from bucardo > > • When doing the pg_dump, I only need to restore it on the > "pg-9.6 slave from bucardo" instance, correct? the bucardo DB does not > store the data? > > Right, “bucardo” database holds meta-information about the syncs, sync > history, etc, but no user data. > > Perfect! > > Thanks! > > > Thanks.
_______________________________________________ Bucardo-general mailing list [email protected] https://mail.endcrypt.com/mailman/listinfo/bucardo-general
