<div dir="ltr">How about follow master command when slave just needs to repoint to the new primary?<div><div>I cannot tell you how many scripts for follow_master_command I have found on the web that take base backup directly or via repmgr. In case of larger database and several slaves this is just plain waste of resources.</div><div>In it's simplest terms follow_master_command shouldn't be nothing more than:</div><div>(1) Stop slave</div><div>(2) Repoint to new master</div><div>(3) Start slave (recover from WAL archive data written by master while being down)</div><div><br></div><div>Is this accurate? </div></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Oct 19, 2018 at 7:01 PM Tatsuo Ishii <<a href="mailto:ishii@sraoss.co.jp">ishii@sraoss.co.jp</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">> It seems that online recovery scripts take base backup of the primary<br>
> to recover downed node.<br>
> With streaming replication and WAL archiving enabled downed node can<br>
> be recovered when it is started again much faster than taking entire<br>
> base backup of the primary.<br>
> So question comes to mind, what is the advantage of using<br>
> time-consuming online recovery over WAL archiving restore process?<br>
> Is attaching downed node to PgPool automatically the only benefit?<br>
<br>
If you are sure that the downed node can be recovered by fetching WAL<br>
from archive at its starting up, you don't need to use online recovery<br>
at all. You just start the standby and issue pcp_attach_node is<br>
enough (assuming that the standby node was once attached to<br>
Pgpool-II).<br>
<br>
On the other hand, if the downed node cannot be recovered due to its<br>
database corruption or some other problems, you have to copy primary<br>
database to standby, and that's what online recovery usually does.<br>
Another scenario is adding brand new standby. In this case you need<br>
fully copy primary database anyway and you should use on line<br>
recovery.<br>
<br>
Best regards,<br>
--<br>
Tatsuo Ishii<br>
SRA OSS, Inc. Japan<br>
English: <a href="http://www.sraoss.co.jp/index_en.php" rel="noreferrer" target="_blank">http://www.sraoss.co.jp/index_en.php</a><br>
Japanese:<a href="http://www.sraoss.co.jp" rel="noreferrer" target="_blank">http://www.sraoss.co.jp</a><br>
</blockquote></div><br clear="all"><div><br></div>-- <br><div dir="ltr" class="gmail_signature"><div dir="ltr"><p><b><span style="color:rgb(64,64,64)">Andre Piwoni</span></b></p>
<p><span style="color:rgb(64,64,64)">Sr. Software Developer,
BI/Database</span></p>
<p><b><i><span style="color:rgb(31,73,125)">Web</span></i></b><span style="color:black">MD Health Services</span><span style="font-family:"Times New Roman",serif;color:black"></span></p>
<p><span style="color:rgb(64,64,64)">Mobile: 801.541.4722</span></p>
<p><span style="color:rgb(31,73,125)"><a href="http://www.webmdhealthservices.com/" target="_blank"><span style="color:blue">www.webmdhealthservices.com</span></a></span><span style="color:rgb(64,64,64)"></span></p></div></div>