Re: gist vacuum seaq access

From: Костя Кузнецов <chapaev28(at)ya(dot)ru>
To: Heikki Linnakangas <hlinnakangas(at)vmware(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: gist vacuum seaq access
Date: 2014-09-13 08:19:43
Message-ID: 3243311410596383@web18h.yandex.ru
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

<p>Heikki.<br />I have idea. when i begining vacuum i must create structure like hash table. And the first action i fill this hash table.<br />in code this look like:<br /><br />for( blkno = ..; all pages; blkno++) {<br />  if(! gistPageIsLeaf(blkno)) {<br />     for( all tuples in blkno ) {<br />         hash[ blkno that is referenced by the tuple] = blkno;<br />     }<br />  }<br />}</p><p>getting is parent of page is fast and simple( hash[page] ).</p><p>But this solution is need 2 full reading of index. </p><p>but Alexander says me that: this solution may be have a problem, because there is <span style="font-size:15px;line-height:21.504px;background-color:#eeeeee;">maintenance_work_mem.</span></p>

Attachment Content-Type Size
unknown_filename text/html 712 bytes

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Fabien COELHO 2014-09-13 08:25:05 Re: pgbench throttling latency limit
Previous Message Ants Aasma 2014-09-13 05:52:33 Re: [REVIEW] Re: Compression of full-page-writes