Stéphane A. Schildknecht wrote:

What you describe might not be a bug.

What should happen in a move set (at a high level) is

1. Slonik contacts the origin and executes moveSet()
2. moveSet() sets all the tables to read only
3. moveSet() creates a MOVE_SET event
4. moveSet() exists -- this is what you see in the log
5. The MOVE_SET event travels to subscribers through replication paths 
and these subscribers reconfigure themselves.  The new origin will 
reconfigure itself when it receives the MOVE_SET command as well.

The set isn't 'moved' until the move set command is confirmed by all nodes.




> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hi,
> 
> I tried to do a move set between two nodes, and I got something strange... All
> my nodes became read only, having write-denying triggers sets on all tables.
> At first glance, I could not see any error in slon logs.
> In PG'log on the old master, I can see : LOG:  duration: 4262.644 ms
> statement: select "_cluster".moveSet(1, 1);
> Nothing on the supposed new master.
> 
> I'm using slony 1.2.16 on postgreSQL 8.3, so I wanted to know if that was a
> known bug which may have been corrected since then, or if I should investigate
> more deeply.
> 
> I'm aware I have to update slon, and knowing this could be the cause of the
> misfunction discovered will certainly help updating :-)
> 
> Thnaks in advance.
> 
> Best regards,
> - -- 
> Stéphane Schildknecht
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.10 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
> 
> iEYEARECAAYFAkxrpHIACgkQA+REPKWGI0GswACgiVpqWmtU1egdd4K1WJHrBi6O
> jZwAn2gNvqBjj+8Rn9ik9wVZIwrg0DBo
> =+8fi
> -----END PGP SIGNATURE-----
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at lists.slony.info
> http://lists.slony.info/mailman/listinfo/slony1-general


-- 
Steve Singer
Afilias Canada
Data Services Developer
416-673-1142