Re: libpq connectoin redirect

<pre style="margin: 0em;">Hi, John,<br /><br />The change will be on the libpq client side.  I am not saying this is a
general<br/>solution for the distributed transaction/scale out.  However, in many cases, it<br />is very useful.  For
example,in my case, I have about 100 departments each has <br />it own database.  The balance machine can just redirect
tothe right box according<br />to database/user.  The 4 boxes I have may not even get domain name or static IP.<br
/>Anotherscenario, if I have some kind of replication set up, I can send transaction<br />processing to master and
analyticreporting query to slaves. <br /><br />Thanks,<br />Feng<br /><br />feng tian wrote:<br /></pre><blockquote
style="border-left:0.2em solid rgb(85, 85, 238); margin: 0em; padding-left: 0.85em;"><pre style="margin: 0em;">Hi,<br
/><br/></pre><tt>I want to load balance a postgres server on 4 physical machines, say </tt><tt>127.0.0.11-14. I can set
upa pgbouncer on 127.0.0.10 and connection </tt><tt>pooling to my four boxes. However, the traffic from/to clients will
</tt><tt>gothrough an extra hop. Another way to do this, is to send the </tt><tt>client an "redirect" message. When
clientconnect to 127.0.0.10, </tt><tt>instead of accepting the connection, it can reply to client telling it
</tt><tt>toreconnect to one of the server on 127.0.0.11-14. </tt><tt> </tt><tt>I am planning to write/submit a patch to
dothat. I wonder if there </tt><tt>is similar effort in extending libpq protocol, or, if you have better </tt><tt>ideas
onhow to achieve this. </tt></blockquote><tt><br />how do you plan on maintaining consistency, transactional integrity
and</tt><tt>atomicity of updates across these 4 machines? </tt><br /><hr />The New Busy is not the too busy. Combine
allyour e-mail accounts with Hotmail. <a
href="http://www.windowslive.com/campaign/thenewbusy?tile=multiaccount&ocid=PID28326::T:WLMTAGL:ON:WL:en-US:WM_HMP:042010_4"
target="_new">Getbusy.</a>