Tue Oct 28 07:58:59 PDT 2008
- Previous message: [Slony1-general] Removing old schemas
- Next message: [Slony1-general] excluding slony from output of pg_dump
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Anoop Bhat <ABhat at trustwave.com> writes: > I've got a database that I was replicating manually (without the altperl scripts) for some time but now I'm using the altperl scripts for replication which obviously make things a lot easier. > > Four questions from all this. > > > 1. if I do a slon_kill -config, how can I restart the replication? Can I just a slon_start -config followed by a slonik_subscribe_set -config set1 node2? You don't need to resubscribe; you merely need to restart the slon processes. If there are updates outstanding, they will be carried over to the subscribers. There are some ways in which Slony-I is pretty fragile; it's not so fragile that killing a slon process breaks it :-). > 2. I want to get rid of all the previous attempts at > replication. Can I simply drop the schemas that were created > previously by me on the master? Slony-I messes (until v2.0) with the internal catalog, so the safe ways to handle this would include: - UNSUBSCRIBE SET To remove one subscriber - DROP NODE/UNINSTALL NODE To clean out one node > 3. what is a good way to check the status of the replication to > see how much is replicated and or what has been replicated thus > far? There is a view in the Slony-I schema called "sl_status." Query that on the origin node, and you'll get a good idea as to how far the subscribers are behind. > 4. I setup all the replication from the slave itself using this as > a rough guide > http://trac.browsershots.org/wiki/DatabaseReplication. Is this > appropriate/recommended or does something need to be initiated from > the master? Previously, I would run a slon process on the master > and a slon on the slave but this seems much easier. Almost too good > to be true. Where you run the slon processes and where you run configuration scripts is your choice; in our environment, we have devoted database servers, as well as "utility servers", and we run slon processes and slonik configuration from the "utility servers." There are good things about running everything from one spot, as that means: - You don't have scripts (slonik and otherwise) strewn across a bunch of servers - You have a single view of all the slon processes, and can be sure they are all running or stopped without needing to ssh to other hosts > Lastly, I just wanted to say that the altperl scripts are so much > easier to deal with and they do seem to make life a lot easier. If only we had someone doing more maintenance on them... -- (format nil "~S@~S" "cbbrowne" "linuxfinances.info") http://cbbrowne.com/info/internet.html Rules of the Evil Overlord #60. "My five-year-old child advisor will also be asked to decipher any code I am thinking of using. If he breaks the code in under 30 seconds, it will not be used. Note: this also applies to passwords." <http://www.eviloverlord.com/>
- Previous message: [Slony1-general] Removing old schemas
- Next message: [Slony1-general] excluding slony from output of pg_dump
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Slony1-general mailing list