<div class="gmail_quote">On Mon, Sep 10, 2012 at 1:26 PM, "Stéphane A. Schildknecht" <span dir="ltr"><<a href="mailto:stephane.schildknecht@postgresql.fr" target="_blank">stephane.schildknecht@postgresql.fr</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div id=":tc">I wondered if there was a magic plan I could rely on to ease my multiple node<br>
replication cluster.<br>
<br>
As there isn't any magical plan, I will have to consider in more deep details<br>
the idea of a two phase migration, as proposed by Christopher.<br></div></blockquote></div><br><div>I think the magic would be "go back in time and do not wait so long between major version upgrades."</div><div>
<br></div><div>It certainly will be safest to do the multi-step upgrade. I'd guess you don't need to do multiple steps for every node in your cluster depending on how they are used...</div><div><br></div>