Fri Jan 28 22:59:12 PST 2005
- Previous message: [Slony1-general] Failover
- Next message: [Slony1-general] Failover
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Fiel Cabral wrote: >My question is about this portion of slony-I-failover.txt: > > >>After failover, getting back node1 >> >> > > > >> After the above failover, the data stored on node1 must be >> considered out of sync with the rest of the nodes. Therefore, the >> only way to get node1 back and transfer the master role to it is >> to rebuild it from scratch as a slave, let it catch up and then >> follow the switchover procedure. >> >> > >Is it necesssary to pg_dump the master's DB to the slave's DB in the >process of "rebuilding it from scratch as a slave"? > >Or is it possible to not do the pg_dump and just "let it catch up"? > > Consider the situation of a table that you never update, but rather only ever insert to. (This is common for transaction tables.) If no "dump" of the data from the provider takes place, and all you do is to add in the changes, that will NEVER include any of the older records. What is done is in fact not a pg_dump, but rather, internally, a "COPY_SET" event. COPY_SET goes through each of the tables that are being replicated in the set, and uses 'COPY table to stdout' to dump the contents of each table. That's kind of analagous to a pg_dump, and there is no choice but to do something reasonably similar to that if you actually want a consistent replica. If you don't care for a consistent replica, I'm not sure what to offer you...
- Previous message: [Slony1-general] Failover
- Next message: [Slony1-general] Failover
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-general mailing list