Hi guys

Replication has died on my master-slave.

What i would really like to do is stop replication and then get it to clear out the replicated tables in the Slave and start again (i.e truncate the replicated tables ONLY ).

The problem I have is that only SOME tables are replicated between the Master and the Slave. 

- i do NOT want the replication to destroy the un-replicated tables on the Slave.

I have seen all the instructions on how to start replication but most of them have the step of drop the Slave db and re-import the schema from the Master. I can't do this as I don't want to clobber my un-replicated tables.

Is it possible to do what I want?

My initial thoughts were to do an slonik "unsubscribe set" and then do a slonik "subscribe set"

The problem I have is that the slon daemon on the Slave is erroring (see below), and from what I understand the "unsubscribe set" event will just be put in the event queue. I.e. The Slave will never be able to unsubscribe because it is erroring on an earlier event.

I was thinking there might be a way to delete all events in the queue so it could just continue ... is this possible? I realise that after that the data is not in sync, but im planning on blowing away the replicated tables and starting again anyway...

Although there might be a way to fix replication, i kind of want to know how to just do a "clean replication" from this point, as I see it as an important task.

Thanks for any thoughts/opinions.

Cheers

Jon

----------------------------------------------------------------



Error in Slave log is:
2013-03-06 19:56:41 HKTINFO   remoteWorkerThread_1: syncing set 1 with 47 table(s) from provider 1
2013-03-06 19:56:42 HKTERROR  remoteWorkerThread_1: "select "_replication".sequenceSetValue(10,1,'5002597503','291048'); " PGRES_FATAL_ERROR ERROR:  Slony-I: sequenceSetValue(): sequence 10 not found