<p dir="ltr">Hi,</p>
<p dir="ltr">Le 23 févr. 2015 16:51, &quot;Clement Thomas&quot; &lt;<a href="mailto:clement1289@gmail.com">clement1289@gmail.com</a>&gt; a écrit :<br>
&gt;<br>
&gt; Hi All,<br>
&gt;           we face a weird problem in our 3 node slony setup.<br>
&gt;<br>
&gt; * node1 (db1.domain.tld )  is the master provider and node2<br>
&gt; (db2.domain.tld ), node3  (db3.domain.tld ) are subscribers.<br>
&gt; currently nodes have 5 replication sets and the replication is working<br>
&gt; fine.<br>
&gt; * the problem is sl_log_1 and sl_log_2 tables in node1 gets cleaned up<br>
&gt; properly, but the tables in the node2 and node3 doesn&#39;t.  On node1 the<br>
&gt; total number of rows in sl_log_1 table is 24845 and in sl_log_2 it is<br>
&gt; 0. whereas<br>
&gt;<br>
&gt; node2:<br>
&gt;<br>
&gt;                          relation                         |  size<br>
&gt; ----------------------------------------------------------+---------<br>
&gt;  _mhb_replication.sl_log_2                                | 130 GB<br>
&gt;  _mhb_replication.sl_log_2_idx1                           | 47 GB<br>
&gt;  _mhb_replication.PartInd_mhb_replication_sl_log_2-node-1 | 30 GB<br>
&gt;<br>
&gt; node3:<br>
&gt;                          relation                         |  size<br>
&gt; ----------------------------------------------------------+--------<br>
&gt;  _mhb_replication.sl_log_2                                | 133 GB<br>
&gt;  _mhb_replication.sl_log_2_idx1                           | 47 GB<br>
&gt;  _mhb_replication.PartInd_mhb_replication_sl_log_2-node-1 | 30 GB<br>
&gt;  _mhb_replication.sl_log_1                                | 352 MB<br>
&gt;<br>
&gt;<br>
&gt; in node2 and node3 could see the following lines frequently.<br>
&gt;<br>
&gt; slon[20695]: [4031-1] FATAL  cleanupThread: &quot;delete from<br>
&gt; &quot;_mhb_replication&quot;.sl_log_1 where log_origin = &#39;1&#39; and log_xid &lt;<br>
&gt; &#39;2130551154&#39;; delete from<br>
&gt; slon[20695]: [4031-2]  &quot;_mhb_replication&quot;.sl_log_2 where log_origin =<br>
&gt; &#39;1&#39; and log_xid &lt; &#39;2130551154&#39;; delete from<br>
&gt; &quot;_mhb_replication&quot;.sl_seqlog where<br>
&gt; slon[20695]: [4031-3]  seql_origin = &#39;1&#39; and seql_ev_seqno &lt;<br>
&gt; &#39;51449379&#39;; select &quot;_mhb_replication&quot;.logswitch_finish(); &quot; - ERROR:<br>
&gt; canceling statement<br>
&gt; slon[20695]: [4031-4]  due to statement timeout<br>
&gt; slon[20695]: [4032-1] DEBUG2 slon_retry() from pid=20695<br>
&gt;<br>
&gt; please find the slony_tools.conf here<br>
&gt; <a href="https://gist.github.com/clement1289/d928acb771ca01a89281">https://gist.github.com/clement1289/d928acb771ca01a89281</a> and sl_status<br>
&gt; /sl_listen output here<br>
&gt; <a href="https://gist.github.com/clement1289/88df40f77c03c691eee5">https://gist.github.com/clement1289/88df40f77c03c691eee5</a> . Hoping for<br>
&gt; some help.<br>
&gt;</p>
<p dir="ltr">You should change PostgreSQL configuration. Statement_timeout is too low. Unset it or set it higher.</p>