Thu Aug 26 20:47:45 PDT 2004
- Previous message: [Slony1-general] Schema Changes
- Next message: [Slony1-general] migrate from 7.4 to 7.5 with slony
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
On August 26, 2004 01:37 pm, joe wrote: > Hi all, > I am testing Slony-I before (hopefully) using it to build a > simple master-slave cluster to run our website. In the documentation > and through various googles I have found that slony doesn't > automagically replicate schema changes. That by itself is not an issue > for me since schema changes don't happen all that often to the website, > but when they do happen it would be nice to know what the proper > sequence of events need to take place to ensure that slony function > properly after, and I haven't been able to determine that sequence of > events is. > I have done some playing with the pg_bench db as setup in the > examples, and what I have found is this: > 1. Alter tables don't seem to require any interaction with > slonik to be replicated, as long as the alter table statement(s) is/are > executed on all nodes before any queries (inserts, updates, deletes) > that use the new schema are run. This doesn't feel quite safe (or sane) > to me and I was wondering how other people handle modifications to > existing tables. You use the execute script directive http://gborg.postgresql.org/project/slony1/genpage.php?slonik_commands#stmt_ddl_script in slonik to accomplish this. (We should probably add a how-to/example of this in action.. I'll look into this mid september if soemone hasn't beaten me too it) > 2. adding new tables to an existing replication set (apparantly) > requires unsubscribing that set, adding the table then re-subscribing > the set. This forces a full dump of the table (and possibly full set?) This is done using merge set http://gborg.postgresql.org/project/slony1/genpage.php?slonik_commands#stmt_merge_set in slonik. (once again more examples of this in practice is probably not a bad thing) > when it happens. (Not fun when the table/set is several gigs). Do most > people just make a new table set when they add tables to the > source/master node or how do they deal with that situation? -- Darcy Buskermolen Wavefire Technologies Corp. ph: 250.717.0200 fx: 250.763.1759 http://www.wavefire.com
- Previous message: [Slony1-general] Schema Changes
- Next message: [Slony1-general] migrate from 7.4 to 7.5 with slony
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-general mailing list