Drop Down MenusCSS Drop Down MenuPure CSS Dropdown Menu

PostgreSQL Database Maintenance Operation

-
Actively updated database tables in PostgreSQL sometimes require maintenance because of table and index bloat, obsolete and suboptimal query plans and sometimes corrupted indexes. Bloated tables and indexes occupy too much of disk space and memory, suboptimal queries and corrupted indexes slow down database queries. As a result, database performance may significantly degrade. PostgreSQL has a few commands to help developers solve those problems: ANALYZE, VACUUM, REINDEX.

ANALYZE

ANALYZE collects statistics about the contents of all or several columns in a table or entire database. These statistics are then used by the query planner to produce efficient execution plans for queries.
Examples:
  • ANALYZE users; collects statistics for users table.
  • ANALYZE VERBOSE users; does exactly the same plus prints progress messages.
  • ANALYZE users (id, display_name); collects statistics for id and display_namecolumns of users table.
  • ANALYZE; collects statistics for all the database tables.
Please note, that ANALYZE doesn't read or update indexes. It deals only with table/column contents. ANALYZE. Other queries may read from the table while ANALYZE is running.
Use ANALYZE when the contents of a table has changed significantly: a few percents of records in a table have been added, updated or deleted. It won't hurt to run ANALYZE just before or right after adding an index to a table. That sometimes helps query planner to properly use the new index.

VACUUM

VACUUM reclaims storage occupied by dead records. Dead records in a table are the deleted or the previous versions of updated records. Such records are only marked as unavailable but still occupy disk space, making table scans slower. VACUUM also has an option that tells it to ANALYZE the table.
There are two very different VACUUM versions. Just VACUUM command reorganises live records in a table and makes the allocated storage space available for new records. However, this storage space doesn't return to the operating system. This command can be run in parallel with other read and write operations on the table. VACUUM FULL exclusively locks the table, creates a new table file, copies only live records to this file, then deletes the old table file. The reclaimed storage space is returned to the system in this case. This command doesn't allow any other read and write operations on the vacuumed table.
Examples:
  • VACUUM users; vacuums users table.
  • VACUUM VERBOSE users; vacuums users table and prints progress messages.
  • VACUUM FULL users; fully vacuums users table. Other queries cannot access userstable while vacuuming is running.
  • VACUUM ANALYZE users; vacuums and analyses users table.
  • VACUUM FULL VERBOSE ANALYZE users; fully vacuums users table and displays progress messages.
  • VACUUM; vacuums all the tables in the database the current user has access to.
Even though PostgreSQL can autovacuum tables itself after a certain number of rows gets marked as deleted, some developers prefer to run VACUUM ANALYZE on tables with a lot of read/write operations on a custom schedule. For example, at the least busy time of the day.
Use VACUUM when the contents of a table has changed significantly: a few percents of records in a table have been added, updated or deleted.

REINDEX

REINDEX rebuilds an existing index or all indexes for a table or entire database. This command is helpful when an index becomes corrupted and is not used anymore, or when an index gets bloated after significant changes in the table contents.
Examples:
  • REINDEX INDEX idx_users_on_display_name; reindexes indexidx_users_on_display_name.
  • REINDEX TABLE users; reindexes users table.
  • REINDEX VERBOSE TABLE users; reindexes users table and prints progress messages.
  • REINDEX DATABASE app_production; reindexes all tables in app_production database.
In my opinion, running VACUUM ANALYZE and then REINDEX commands on the biggest tables in the application database really helps to improve its performance.

Comments

Post a Comment

Popular posts from this blog

postgreSQL Compress format backup

-
The -F option on pg_dump allows you to specify the format of the output. The default output is plain-text. That’s what you get when you do this: pg_dump bacula > bacula.sql and is functionally equivalent to doing: pg_dump -Fp bacula > bacula.sql There are other formats (directory, tar), but today I’m concerned with the custom (c) option. By default, this output is also compressed (see the -Z option). What is of most value is the ability to use  pg_restore  with the custom output. What’s the big deal about custom format? The more interesting option on pg_restore was -j. This allows you to run multiple concurrent jobs for the loading of data and building of indexes. This can greatly reduce the amount of time needed to restore a large database if your server has multiple processors. pg_restore can be much more useful than just doing: psql bacula < bacula.sql With pg_restore you can be very selective about what you restore. See the -L and -l options. How much do
Read more

How to find the server is whether standby (slave) or primary(master) in Postgresql replication ?

-
Method 1 You can check the mode of the server using "pg_controldata". [pgsql@test~]$ pg_controldata /usr/local/pgsql/data84/ Database cluster state: in archive recovery --> This is Standby Database Database cluster state: in production --> This is Production Database [Master] Method 2 You can use pg_is_in_recovery() which returns True if recovery is still in progress(so the server is running in standby mode or slave) postgres=# select pg_is_in_recovery(); pg_is_in_recovery ------------------- t (1 row) If Return    false   so the server is running in primary mode or master postgres=# select pg_is_in_recovery(); pg_is_in_recovery ------------------- f (1 row)
Read more

Top 10 Highest Paying URL Shortener Companies - Best URL Shorteners To Earn Fast

-
Image
These are the lists of top paying url shortening websites that are similar to Adfoc.us or Shrink url Websites like Adfoc.us. The websites are similar in the sense that you can use the same pattern of registration and earnings for all them, but the only difference is the amount of money each pays to its members and also the percentage of earnings from their referral programs. 1. ADF.LY – Free Url Shortening similar to adfocus that will pay you $4 for 1000 clicks on your shorten url. You can Also refer your friends and earn 20% of their earnings for life. Their payment is quick and consistant. Paypal Or Alertpay. It’s legit and very popular. JOIN  HERE Payment Options Details :- Payout for 1000 Views: Upto $9 Minimum Payout: $5 Referral Commission: 20% Payment Options:Payoneer, Payza Payment Time : Daily and On Request You can Withdraw money Following Accounts:- 2. LINKBUCKS – free url shortening LINKBUCKS is a free url shortening service website si
Read more

7 Steps to configure BDR replication in postgresql

-
Image
 The  BDR  (Bi-Directional Replication) project adds multi-master replication to PostgreSQL 9.4.  Postgres-BDR has a lower impact on the masters(s) than trigger-based replication solutions. There is no write-amplification, as it does not require triggers to write to queue tables in order to replicate writes.   Here We are using  postgres version 9.4.12 and bdr version 1.0.2. for configuring  multi master replication . Simply Following 7 steps you can configure the multi master replication in postgresql. To download the bdr in below link. .   https://github.com/2ndQuadrant/bdr/archive/bdr-pg/REL9_4_12-1.tar.gz $ tar -xzvf REL9_4_12-1.tar.gz $ wget https://github.com/2ndQuadrant/bdr/archive/bdr-plugin/1.0.2.tar.gz $ tar -xzvf 1.0.2.tar.gz 1. To install BDR. $ cd ~/bdr-bdr-pg-REL9_4_12-1 $ ./configure --prefix=/usr/lib/postgresql/9.4 --enable-debug --with-openssl $ make -j4 -s install-world $ cd ~/bdr-bdr-plugin-1.0.2 $ PATH=/usr/lib/postgresql/9.4/bin:"$PATH" ./c
Read more

Postgres Streaming Replication Configuration

-
Primary - Existing Test Server: ======= Hostname: nijam-1 IP Address: 53.99.198.11 Secodary - New Test Server ======== Hostname: nijam-2 IP Address: 53.99.198.12 ===================================================== Step 1. stop the slave server /optt/10.5/bin/pg_ctl -D /DATA stop Step 2. move the slave data directory to backup location: mv /DATA/* /tmp/backup Step 3. change below parameter on master postgresql.conf & pg_hba.conf files . listen_addresses = 'localhost,53.99.198.11' wal_level = replica # minimal, replica, or logical archive_mode = on archive_command = 'cp %p /ARCHIVES/%f' max_wal_senders = 10 wal_keep_segments = 25 On pg_hba.conf: host replication postgres 53.99.198.12/24 trust host replication postgres 53.99.198.11/24 trust Step 4. Restart the master server: /optt/10.5/bin/pg_ctl -D /DATA restart Step 5. taking consistancy backup: psql -c "sele
Read more