On 12-10-05 08:37 AM, Mike James wrote:
> Here's what the DBA did as I understand it. This is his procedure for changing the replication set. Slony version is slony1-2.0.3-rc. Postgres 8.3.9 on the Master. I might be mis-using some of the terminology, as I'm not a DBA and this is my 1st foray with slony.
> 1. Kill the slon running on the master and slave. Leave the slon on the remote host running.
> 2. he uses PGAdmin to remove the slony schema (slony replication cluster) from Master and Slave.
> 3. regenerate the set records using a shell script on the Master.
> 4. use slonik to initialize the cluster with the set records from above on Master and Slave.
> 5. start the slon on the Master.
> 6. start the slon on the Slave with -a option.
> 7. use slonik to subscribe the Slave to the Master.
>
> On the remote postgres server, the value of at_counter in the _cluster1.sl_archive_tracking table is 338005. But it looks like the new logs being shipped started over at sequence number 1.
>
> Does that description make sense? This can't be the optimal way to add a table to the replication set!
>
What you describe makes perfect sense.  Step 2, removing the slony 
schema uninstalls slony from both the master and the slave.  When you 
reinstall slony with step 4 you are creating a new slony cluster.  The 
offline node isn't in sync with the new cluster it was sync'd with the 
old one.  You will need to rebuild your offline node with slony1_dump.sh

The proper way to add tables to a replication set is described here
http://www.slony.info/documentation/2.0/administration.html#ADDTHINGS

If your using the altperl tools you can look at slonik_create_set, 
slonik_subscribe_set and slonik_merge_sets



> Mike
>
> -----Original Message-----
> From: Steve Singer [mailto:ssinger at ca.afilias.info]
> Sent: Thursday, October 04, 2012 5:39 PM
> To: Mike James
> Cc: slony1-general at lists.slony.info
> Subject: Re: [Slony1-general] offline log shipping errors
>
> On 12-10-04 04:11 PM, Mike James wrote:
>> OK, I was mistaken. The slon on the subscriber that ships the logs was restarted with the -a option. The slon on the Master was not started with the -a option.
>
> Then I'm a bit confused about steps were done that resulted in log shipping getting out of sync.
>
> 1. If setup log shipping and started slon with -a on the slave 2. Added some tables to replication
>
> I would expect log shipping to continue to work.
>
>
>
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at lists.slony.info
> http://lists.slony.info/mailman/listinfo/slony1-general
>