Re: pre-commit triggers - Mailing list pgsql-hackers

From Hannu Krosing
Subject Re: pre-commit triggers
Date
Msg-id 5287CEF9.3020908@2ndQuadrant.com
Whole thread Raw
In response to pre-commit triggers  (Andrew Dunstan <andrew@dunslane.net>)
Responses Re: pre-commit triggers  (Andrew Dunstan <andrew@dunslane.net>)
Re: pre-commit triggers  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
<div class="moz-cite-prefix">On 11/15/2013 07:01 PM, Andrew Dunstan wrote:<br /></div><blockquote
cite="mid:5286618C.60301@dunslane.net"type="cite"><br /> Attached is a patch to provide a new event trigger that will
fireon transaction commit. I have tried to make certain that it fires at a sufficiently early stage in the commit
processthat some of the evils mentioned in previous discussions on this topic aren't relevant. <br /><br /> The
triggersdon't fire if there is no real XID, so only actual data changes should cause the trigger to fire. <br
/></blockquote>I have not looked at the patch, but does it also run pre-rollback ?<br /><br /> If not, how hard would
itbe to make it so ?<br /><br /><blockquote cite="mid:5286618C.60301@dunslane.net" type="cite">They also don't fire in
singleuser mode, so that if you do something stupid like create a trigger that unconditionally raises an error you have
away to recover. <br /><br /> This is intended to be somewhat similar to the same feature in the Firebird database, and
theinitial demand came from a client migrating from that system to Postgres. <br /><br /> cheers <br /><br /> andrew
<br/><br /><fieldset class="mimeAttachmentHeader"></fieldset><br /><pre wrap="">
 
</pre></blockquote><br /><br /><pre class="moz-signature" cols="72">-- 
Hannu Krosing
PostgreSQL Consultant
Performance, Scalability and High Availability
2ndQuadrant Nordic OÜ</pre>

pgsql-hackers by date:

Previous
From: Fujii Masao
Date:
Subject: Re: New option for pg_basebackup, to specify a different directory for pg_xlog
Next
From: Миша Тюрин
Date:
Subject: freeze cannot be finished