<p dir="ltr"><br>
On 28 Apr 2014 19:29, "Jan Wieck" <<a href="mailto:jan@wi3ck.info">jan@wi3ck.info</a>> wrote:<br>
><br>
> On 04/28/14 08:01, Steve Singer wrote:<br>
>></p>
<p dir="ltr">Thank you Steve/Jan.</p>
<p dir="ltr">As I said, I was just trying to prevent duplicate row in that way. Not that it will completely take care of it.</p>
<p dir="ltr"> <br>
> since the index isn't restricted to the primary key columns of the<br>
> tAble. And since it is shared by all user tables, that's impossible.<br>
></p>
<p dir="ltr">How about capturing primary key value in catalogs like its done for relation OID?<br>
(Its a last guess :) )</p>
<p dir="ltr">> 3) It will also no t prevent something like this:<br>
><br>
><br>
> insert into dtest values (1, 'D');<br>
> -- wait for 30 minutes in which slony cleanup switches logs and<br>
> -- the index corruption happens<br>
><br>
> insert into dtest values (1, 'D');<br>
></p>
<p dir="ltr">This is the next case i was thinking , it becomes difficult in my case too to find duplicate row in sl_log_N.</p>
<p dir="ltr">I belive this type of scenario should be described somewhere in slony faq for user reference.</p>
<p dir="ltr">-Raghav <br>
><br>
> Regards,<br>
> Jan<br>
><br>
><br>
><br>
> -- <br>
> Jan Wieck<br>
> Senior Software Engineer<br>
> <a href="http://slony.info">http://slony.info</a><br>
</p>