On 8/9/2010 10:55 AM, Steve Singer wrote:
> Attached are a set of patches for addressing the issues discussed in bug136.
> 
> These issues show up when you have a forwarding node that fails. The 
> nodes that were being provided from the failed node need to be 
> reconfigured to get there data elsewhere.  The issue is that even if 
> path exists between the provider nodes and other nodes in the system the 
> provider node might not be listening for these events so it might never 
> no to reconfigure itself.
> 
> This patch set tries to deal with both cases where a FAILOVER is issued 
>   and cases where the failed node is just a forwarder so a SUBSCRIBE SET 
> is issued but no FAILOVER.
> 
> The intent is that these patches will be applied with the ones for 
> bug139 and 132.
> 
> http://github.com/ssinger/slony1-engine/tree/bug139_132_136
> http://github.com/ssinger/slony1-engine/tree/bug136

I think reshapeSubscription() should only initiate a slon restart if it 
actually updated a row in sl_subscribe. Otherwise it will needlessly 
restart a slon on a regular subscription.


Jan

-- 
Anyone who trades liberty for security deserves neither
liberty nor security. -- Benjamin Franklin