Skip site navigation (1) Skip section navigation (2)

Re: SSI SLRU strategy choices

From: "Kevin Grittner" <Kevin(dot)Grittner(at)wicourts(dot)gov>
To: "Alvaro Herrera" <alvherre(at)commandprompt(dot)com>
Cc: "Heikki Linnakangas" <heikki(dot)linnakangas(at)enterprisedb(dot)com>, "pgsql-hackers" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: SSI SLRU strategy choices
Date: 2010-12-29 20:36:55
Message-ID: 4D1B47880200002500038D98@gw.wicourts.gov (view raw or flat)
Thread:
Lists: pgsql-hackers
"Kevin Grittner" <Kevin(dot)Grittner(at)wicourts(dot)gov> wrote:
 
> if a serializable transaction which is not flagged as read only
> remains open long enough for over a billion other transactions to
> commit
 
Maybe a clarification and example would be useful.  We're talking
about going through a billion transactions which were assigned a
TransactionId, not all database transactions.  An example of how you
could hit that is with a sustained commit rate of 5000 transactions
per second which are modifying data while a single read write
transaction stays open for 2.3 days.
 
-Kevin

In response to

pgsql-hackers by date

Next:From: Simon RiggsDate: 2010-12-29 20:58:18
Subject: Re: small table left outer join big table
Previous:From: David E. WheelerDate: 2010-12-29 20:36:44
Subject: Re: Extensions, patch v16

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group