Re: Buglist - Mailing list pgsql-general

From Vivek Khera
Subject Re: Buglist
Date
Msg-id 16195.42545.204240.993936@yertle.int.kciLink.com
Whole thread Raw
In response to Re: Buglist  (Alvaro Herrera <alvherre@dcc.uchile.cl>)
List pgsql-general
>>>>> "AH" == Alvaro Herrera <alvherre@dcc.uchile.cl> writes:

>> once within a vacuum process.  I am fairly certian it would reduce
>> disk bandwidth requirements since at least one necessary page will
>> already be in memory.

AH> There's no way to check for "external references", because said
AH> references are actually the running transactions.  So you can't drop a
AH> row until all the transactions that were active during your transaction

How does the backend running vacuum detect that a row has no active
references?  Why can't the current backend know that same information?

pgsql-general by date:

Previous
From: Vivek Khera
Date:
Subject: Re: Buglist
Next
From: Jan Wieck
Date:
Subject: Re: Buglist