i've never managed to get it to work using pg_dump due to OIDs... I 
always do a physical copy of the postgres data directory when server is 
shut down.. only way to do it reliably and be sure everything is there 
and exactly the same.

using b2zip as i think rsync has some optimisation for the p2zip as 
well.. and it got your bandwidth limiter as well :)


Shahaf Abileah wrote:
> I haven't tried to update the path on a given node before.  That's
> clever.  Do you think it would work to do a pg_dump / pg_restore?  Or
> would that create different OID's and therefore break replication?
>
> We have copied files over using rsync (which has a bwlimit option) and
> that seems to work very well.
>
> --S
>
>
> -----Original Message-----
> From: slony1-general-bounces at lists.slony.info
> [mailto:slony1-general-bounces at lists.slony.info] On Behalf Of Martin
> Eriksson
> Sent: Friday, August 22, 2008 1:06 AM
> Cc: slony1-general at lists.slony.info
> Subject: Re: [Slony1-general] bandwidth limit on slony replication
>
> I don't know of any such ways in slony, there might be?
>
> even if there is no such way in slony you could always throttle it on a 
> lower level on the network level, pleanty of software available to do
> that..
>
> but there are other ways to handle the initial huge bulk load as well..
>
> e.g. replicate to a new instance in your data center (assuming its same 
> hardware architecture (32bit vs 64bit , intel vs AMD etc, postgreSQL 
> version) as the new node) once replication is done shut down the 
> instance and do a physical copy of the data directory of the new 
> instance. if you b2zip it will become about 5-7 times smaller and then 
> transfer that and use that data directory for the new node, modify the 
> slon path with slony "store path" to correspond with the new ip etc. 
> then just let it catch up on the data since the shutdown.
>
> this would yield 4-7 times less bandwidth usage, other way is to put it 
> on a removable disk and bring it physical to the new location.
>
> I've done both ways in our slony cluster and never had any problem. But 
> then again we always make sure all our db machines are exactly the same 
> in terms of hardware and software.
>
> Martin
>
>
> Shahaf Abileah wrote:
>   
>> Is there a way to limit the amount of bandwidth that slony uses?
>>
>> Within our data center we don't really care how much network is used 
>> to replicate DB's from one server to another. However, we'd like to 
>> set up slaves in other locations, and we can get slapped with huge 
>> fees if we have big spikes of network bandwidth usage (e.g. when you 
>> first set up a slave and you need to bulk-copy all the tables).
>>
>> Thanks,
>>
>> --S
>>
>>
>>     
> ------------------------------------------------------------------------
>   
>> _______________________________________________
>> Slony1-general mailing list
>> Slony1-general at lists.slony.info
>> http://lists.slony.info/mailman/listinfo/slony1-general
>>   
>>     
>
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at lists.slony.info
> http://lists.slony.info/mailman/listinfo/slony1-general
>
>
>