Services
24×7×365 Technical Support
Migration to PostgreSQL
High Availability Deployment
Database Audit
Remote DBA for PostgreSQL
Products
Postgres Pro Enterprise
Postgres Pro Standard
Cloud Solutions
Postgres Extensions
Resources
Blog
Documentation
Webinars
Videos
Presentations
Community
Events
Training Courses
Books
Demo Database
Mailing List Archives
About
Leadership team
Partners
Customers
In the News
Press Releases
Press Info
Services
24×7×365 Technical Support
Migration to PostgreSQL
High Availability Deployment
Database Audit
Remote DBA for PostgreSQL
Products
Postgres Pro Enterprise
Postgres Pro Standard
Cloud Solutions
Postgres Extensions
Resources
Blog
Documentation
Webinars
Videos
Presentations
Community
Events
Training Courses
Books
Demo Database
Mailing List Archives
About
Leadership team
Partners
Customers
In the News
Press Releases
Press Info
Facebook
Downloads
Home
>
mailing lists
Zero dead tuples, when significant apparent bloat - Mailing list pgsql-general
From
John Melesky
Subject
Zero dead tuples, when significant apparent bloat
Date
December 11, 2013
04:27:07
Msg-id
CAJ1GNCotqp=9qyCJAwz_20Q769x5jZ5C7AgR-qg8xhaBbyfQUA@mail.gmail.com
Whole thread
Raw
Responses
Re: Zero dead tuples, when significant apparent bloat
(bricklen <bricklen@gmail.com>)
Re: Zero dead tuples, when significant apparent bloat
(Tom Lane <tgl@sss.pgh.pa.us>)
List
pgsql-general
Tree view
Here's the situation:
relation_size | indexes_size | total_relation_size
----------------+----------------+---------------------
997 MB | 2073 MB | 3070 MB
(1 row)
If I select n_dead_tup from pg_stat_user_tables, I get:
n_dead_tup
------------
0
Okay, so I run ANALYZE table, then check again:
n_dead_tup
------------
0
Finally, I run VACUUM FULL ANALYZE table. Now I see:
relation_size | indexes_size | total_relation_size
----------------+----------------+---------------------
118 MB | 200 MB | 319 MB
(1 row)
.... ?
It seems clear that there were dead tuples, since the table size shrank to an eighth of its previous size. Why did analyze not pick that up?
Am I missing something?
This is a very large database, so we want to introspect against live/dead tuple percentage to minimize the tables we run a VACUUM FULL against.
I've been staring at this for days.
-john
--
John Melesky | Sr Database Administrator
503.284.7581 x204
|
john.melesky@rentrak.com
RENTRAK |
www.rentrak.com
| NASDAQ: RENT
Notice: This message is confidential and is intended only for the
recipient(s) named above. If you have received this message in error,
or are not the named recipient(s), please immediately notify the
sender and delete this message.
pgsql-general
by date:
Previous
From:
Jerry Sievers
Date:
11 December 2013, 02:00:08
Subject:
Re: validate synatax
Next
From:
bricklen
Date:
11 December 2013, 04:57:33
Subject:
Re: Zero dead tuples, when significant apparent bloat
Есть вопросы? Напишите нам!
Соглашаюсь с условиями обработки персональных данных
I confirm that I have read and accepted PostgresPro’s
Privacy Policy
.
I agree to get Postgres Pro discount offers and other marketing communications.
✖
×
×
Everywhere
Documentation
Mailing list
List:
all lists
pgsql-general
pgsql-hackers
buildfarm-members
pgadmin-hackers
pgadmin-support
pgsql-admin
pgsql-advocacy
pgsql-announce
pgsql-benchmarks
pgsql-bugs
pgsql-chat
pgsql-cluster-hackers
pgsql-committers
pgsql-cygwin
pgsql-docs
pgsql-hackers-pitr
pgsql-hackers-win32
pgsql-interfaces
pgsql-jdbc
pgsql-jobs
pgsql-novice
pgsql-odbc
pgsql-patches
pgsql-performance
pgsql-php
pgsql-pkg-debian
pgsql-pkg-yum
pgsql-ports
pgsql-rrreviewers
pgsql-ru-general
pgsql-sql
pgsql-students
pgsql-testers
pgsql-translators
pgsql-www
psycopg
Period
anytime
within last day
within last week
within last month
within last 6 months
within last year
Sort by
date
reverse date
rank
Services
24×7×365 Technical Support
Migration to PostgreSQL
High Availability Deployment
Database Audit
Remote DBA for PostgreSQL
Products
Postgres Pro Enterprise
Postgres Pro Standard
Cloud Solutions
Postgres Extensions
Resources
Blog
Documentation
Webinars
Videos
Presentations
Community
Events
Training Courses
Books
Demo Database
Mailing List Archives
About
Leadership team
Partners
Customers
In the News
Press Releases
Press Info
By continuing to browse this website, you agree to the use of cookies. Go to
Privacy Policy
.
I accept cookies