October 20, 2018

Good intentions going bad, Binding Views Unique indexing and noexpand

Related imageYesterday was a very interesting day, trying to tune a procedure that uses a schema binding view, I try to add a needed unique clustered index into the view, so I can utilize the noexpand hint to used the actual indexes from the table, the result a total cluster ckfu. First there is not online support to build indexes on a view so the table got lock when adding the unique clustered index block the table, which cause everyone to jump in panic, while I was unaware of the situation as a let the script execute in a reduce script while working on another task. So lesson learns, monitor the execution of the script, if blocking way until a later time (maintenance window), the indexes will improved the execution to a 45% faster response time, but, careful planning should be in place, sometimes we get wrap into fixing and not realizing that everyone won't remember what you did good, but that one day when everything went to hell.

August 17, 2018

PostgresSQL Scripts References.

A lot of work working on postgresql and today was another day of performance and fixing issues, eventually let add this set of scripts giving by one of my peers (Thanks Troy), now that I have them is time to shared with the world.


Is this a master or read only server I'm on now?  Can it accept writes?


select pg_is_in_recovery();

or, just try to create a database.  You will fail if it's a standby.




Check server activity:


select * from pg_stat_activity;




2 queries for database blocking and locking ( to be run on master node )


SELECT blocked_locks.pid     AS blocked_pid,

blocked_activity.usename  AS blocked_user,

blocking_locks.pid     AS blocking_pid,

blocking_activity.usename AS blocking_user,

blocked_activity.query    AS blocked_statement,

blocking_activity.query   AS current_statement_in_blocking_process

FROM  pg_catalog.pg_locks         blocked_locks

JOIN pg_catalog.pg_stat_activity blocked_activity  ON blocked_activity.pid = blocked_locks.pid

JOIN pg_catalog.pg_locks         blocking_locks

ON blocking_locks.locktype = blocked_locks.locktype


AND blocking_locks.relation IS NOT DISTINCT FROM blocked_locks.relation

AND blocking_locks.page IS NOT DISTINCT FROM blocked_locks.page

AND blocking_locks.tuple IS NOT DISTINCT FROM blocked_locks.tuple

AND blocking_locks.virtualxid IS NOT DISTINCT FROM blocked_locks.virtualxid

AND blocking_locks.transactionid IS NOT DISTINCT FROM blocked_locks.transactionid

AND blocking_locks.classid IS NOT DISTINCT FROM blocked_locks.classid

AND blocking_locks.objid IS NOT DISTINCT FROM blocked_locks.objid

AND blocking_locks.objsubid IS NOT DISTINCT FROM blocked_locks.objsubid

AND blocking_locks.pid != blocked_locks.pid

JOIN pg_catalog.pg_stat_activity blocking_activity ON blocking_activity.pid = blocking_locks.pid

WHERE NOT blocked_locks.GRANTED;


# There's also:



waiting.locktype           AS waiting_locktype,

waiting.relation::regclass AS waiting_table,

waiting_stm.query          AS waiting_query,

waiting.mode               AS waiting_mode,

waiting.pid                AS waiting_pid,

other.locktype             AS other_locktype,

other.relation::regclass   AS other_table,

other_stm.query            AS other_query,

other.mode                 AS other_mode,

other.pid                  AS other_pid,

other.GRANTED              AS other_granted


pg_catalog.pg_locks AS waiting


pg_catalog.pg_stat_activity AS waiting_stm

ON (waiting_stm.pid = waiting.pid)


pg_catalog.pg_locks AS other

ON ((waiting."database" = other."database"

AND waiting.relation  = other.relation

)OR waiting.transactionid = other.transactionid)

JOIN pg_catalog.pg_stat_activity AS other_stm

ON (other_stm.pid = other.pid)


AND waiting.pid <> other.pid;




# Recursive view of blocking



     c(requested, CURRENT) AS

       ( VALUES

         ('AccessShareLock'::text, 'AccessExclusiveLock'::text),

         ('RowShareLock'::text, 'ExclusiveLock'::text),

         ('RowShareLock'::text, 'AccessExclusiveLock'::text),

         ('RowExclusiveLock'::text, 'ShareLock'::text),

         ('RowExclusiveLock'::text, 'ShareRowExclusiveLock'::text),

         ('RowExclusiveLock'::text, 'ExclusiveLock'::text),

         ('RowExclusiveLock'::text, 'AccessExclusiveLock'::text),

         ('ShareUpdateExclusiveLock'::text, 'ShareUpdateExclusiveLock'::text),

         ('ShareUpdateExclusiveLock'::text, 'ShareLock'::text),

         ('ShareUpdateExclusiveLock'::text, 'ShareRowExclusiveLock'::text),

         ('ShareUpdateExclusiveLock'::text, 'ExclusiveLock'::text),

         ('ShareUpdateExclusiveLock'::text, 'AccessExclusiveLock'::text),

         ('ShareLock'::text, 'RowExclusiveLock'::text),

         ('ShareLock'::text, 'ShareUpdateExclusiveLock'::text),

         ('ShareLock'::text, 'ShareRowExclusiveLock'::text),

         ('ShareLock'::text, 'ExclusiveLock'::text),

         ('ShareLock'::text, 'AccessExclusiveLock'::text),

         ('ShareRowExclusiveLock'::text, 'RowExclusiveLock'::text),

         ('ShareRowExclusiveLock'::text, 'ShareUpdateExclusiveLock'::text),

         ('ShareRowExclusiveLock'::text, 'ShareLock'::text),

         ('ShareRowExclusiveLock'::text, 'ShareRowExclusiveLock'::text),

         ('ShareRowExclusiveLock'::text, 'ExclusiveLock'::text),

         ('ShareRowExclusiveLock'::text, 'AccessExclusiveLock'::text),

         ('ExclusiveLock'::text, 'RowShareLock'::text),

         ('ExclusiveLock'::text, 'RowExclusiveLock'::text),

         ('ExclusiveLock'::text, 'ShareUpdateExclusiveLock'::text),

         ('ExclusiveLock'::text, 'ShareLock'::text),

         ('ExclusiveLock'::text, 'ShareRowExclusiveLock'::text),

         ('ExclusiveLock'::text, 'ExclusiveLock'::text),

         ('ExclusiveLock'::text, 'AccessExclusiveLock'::text),

         ('AccessExclusiveLock'::text, 'AccessShareLock'::text),

         ('AccessExclusiveLock'::text, 'RowShareLock'::text),

         ('AccessExclusiveLock'::text, 'RowExclusiveLock'::text),

         ('AccessExclusiveLock'::text, 'ShareUpdateExclusiveLock'::text),

         ('AccessExclusiveLock'::text, 'ShareLock'::text),

         ('AccessExclusiveLock'::text, 'ShareRowExclusiveLock'::text),

         ('AccessExclusiveLock'::text, 'ExclusiveLock'::text),

         ('AccessExclusiveLock'::text, 'AccessExclusiveLock'::text)


     l AS



             (locktype,DATABASE,relation::regclass::text,page,tuple,virtualxid,transactionid,classid,objid,objsubid) AS target,





           FROM pg_catalog.pg_locks


     t AS



             blocker.target  AS blocker_target,

             blocker.pid     AS blocker_pid,

             blocker.mode    AS blocker_mode,

             blocked.target  AS target,

             blocked.pid     AS pid,

             blocked.mode    AS mode

           FROM l blocker

           JOIN l blocked

             ON ( NOT blocked.GRANTED

              AND blocker.GRANTED

              AND blocked.pid != blocker.pid

              AND blocked.target IS NOT DISTINCT FROM blocker.target)

           JOIN c ON (c.requested = blocked.mode AND c.CURRENT = blocker.mode)


     r AS






             '1'::INT        AS depth,




             blocker_pid::text || ',' || pid::text AS seq

           FROM t

         UNION ALL





             blocker.depth + 1,




             blocker.seq || ',' || blocked.pid::text

           FROM r blocker

           JOIN t blocked

             ON (blocked.blocker_pid = blocker.pid)

           WHERE blocker.depth < 1000



  ORDER BY seq;




Need more info that the logs aren't giving you?  Adjust the postgresql.conf file to allow verbose logging.

You'll need to load the config changes by "pg_ctl reload", then check the logs to ensure they were loaded.


log_duration = on|off

log_lock_waits = on|off

log_min_duration_statement = Xs (can be adjusted, in seconds)

deadlock_timeout = Xs (can be adjusted, in seconds)




pg_buffercache (create extension pg_buffercache;) to create the extension if it isn't already there.


SELECT c.relname, count(*) AS buffers

FROM pg_buffercache b INNER JOIN pg_class c

ON b.relfilenode = pg_relation_filenode(c.oid) AND

b.reldatabase IN (0, (SELECT oid FROM pg_database

WHERE datname = current_database()))

GROUP BY c.relname


LIMIT 100;




postgres cache hit ratio (create extension pg_buffercache;) to create the extension if it isn't already there.


SELECT sum(heap_blks_read) as heap_read,

sum(heap_blks_hit)  as heap_hit,

sum(heap_blks_hit) / (sum(heap_blks_hit) + sum(heap_blks_read)) as ratio

FROM pg_statio_user_tables;




postgres index cache hit ratio (create extension pg_buffercache;) to create the extension if it isn't already there.


SELECT sum(idx_blks_read) as idx_read,

sum(idx_blks_hit)  as idx_hit,

(sum(idx_blks_hit) - sum(idx_blks_read)) / sum(idx_blks_hit) as ratio

FROM pg_statio_user_indexes;



Is postgres using ssl for its connections?  Let's see:


select datname, usename, client_addr, ssl, cipher

from pg_stat_activity

join pg_stat_ssl on pg_stat_activity.pid=pg_stat_ssl.pid;





Table and index stats


with table_stats as (

select psut.relname,


  1.0 * psut.idx_scan / greatest(1, psut.seq_scan + psut.idx_scan) as index_use_ratio

from pg_stat_user_tables psut

order by psut.n_live_tup desc


table_io as (

select psiut.relname,

  sum(psiut.heap_blks_read) as table_page_read,

  sum(psiut.heap_blks_hit)  as table_page_hit,

  sum(psiut.heap_blks_hit) / greatest(1, sum(psiut.heap_blks_hit) + sum(psiut.heap_blks_read)) as table_hit_ratio

from pg_statio_user_tables psiut

group by psiut.relname

order by table_page_read desc


index_io as (

select psiui.relname,


  sum(psiui.idx_blks_read) as idx_page_read,

  sum(psiui.idx_blks_hit) as idx_page_hit,

  1.0 * sum(psiui.idx_blks_hit) / greatest(1.0, sum(psiui.idx_blks_hit) + sum(psiui.idx_blks_read)) as idx_hit_ratio

from pg_statio_user_indexes psiui

group by psiui.relname, psiui.indexrelname

order by sum(psiui.idx_blks_read) desc


select ts.relname, ts.n_live_tup, ts.index_use_ratio,

  ti.table_page_read, ti.table_page_hit, ti.table_hit_ratio,

  ii.indexrelname, ii.idx_page_read, ii.idx_page_hit, ii.idx_hit_ratio

from table_stats ts

left outer join table_io ti

  on ti.relname = ts.relname

left outer join index_io ii

  on ii.relname = ts.relname

order by ti.table_page_read desc, ii.idx_page_read desc





Table & Index Bloat query (from check_postgres script)



  current_database(), schemaname, tablename, /*reltuples::bigint, relpages::bigint, otta,*/

  ROUND((CASE WHEN otta=0 THEN 0.0 ELSE sml.relpages::FLOAT/otta END)::NUMERIC,1) AS tbloat,

  CASE WHEN relpages < otta THEN 0 ELSE bs*(sml.relpages-otta)::BIGINT END AS wastedbytes,

  iname, /*ituples::bigint, ipages::bigint, iotta,*/

  ROUND((CASE WHEN iotta=0 OR ipages=0 THEN 0.0 ELSE ipages::FLOAT/iotta END)::NUMERIC,1) AS ibloat,

  CASE WHEN ipages < iotta THEN 0 ELSE bs*(ipages-iotta) END AS wastedibytes



    schemaname, tablename, cc.reltuples, cc.relpages, bs,


      (CASE WHEN datahdr%ma=0 THEN ma ELSE datahdr%ma END))+nullhdr2+4))/(bs-20::FLOAT)) AS otta,

    COALESCE(c2.relname,'?') AS iname, COALESCE(c2.reltuples,0) AS ituples, COALESCE(c2.relpages,0) AS ipages,

    COALESCE(CEIL((c2.reltuples*(datahdr-12))/(bs-20::FLOAT)),0) AS iotta -- very rough approximation, assumes all cols

  FROM (



      (datawidth+(hdr+ma-(CASE WHEN hdr%ma=0 THEN ma ELSE hdr%ma END)))::NUMERIC AS datahdr,

      (maxfracsum*(nullhdr+ma-(CASE WHEN nullhdr%ma=0 THEN ma ELSE nullhdr%ma END))) AS nullhdr2

    FROM (


        schemaname, tablename, hdr, ma, bs,

        SUM((1-null_frac)*avg_width) AS datawidth,

        MAX(null_frac) AS maxfracsum,


          SELECT 1+COUNT(*)/8

          FROM pg_stats s2

          WHERE null_frac<>0 AND s2.schemaname = s.schemaname AND s2.tablename = s.tablename

        ) AS nullhdr

      FROM pg_stats s, (


          (SELECT current_setting('block_size')::NUMERIC) AS bs,

          CASE WHEN SUBSTRING(v,12,3) IN ('8.0','8.1','8.2') THEN 27 ELSE 23 END AS hdr,

          CASE WHEN v ~ 'mingw32' THEN 8 ELSE 4 END AS ma

        FROM (SELECT version() AS v) AS foo

      ) AS constants

      GROUP BY 1,2,3,4,5

    ) AS foo

  ) AS rs

  JOIN pg_class cc ON cc.relname = rs.tablename

  JOIN pg_namespace nn ON cc.relnamespace = nn.oid AND nn.nspname = rs.schemaname AND nn.nspname <> 'information_schema'

  LEFT JOIN pg_index i ON indrelid = cc.oid

  LEFT JOIN pg_class c2 ON c2.oid = i.indexrelid

) AS sml

ORDER BY wastedbytes DESC;




relations buffered in database share buffer


select c.relname,pg_size_pretty(count(*) * 8192) as buffered,

        round(100.0 * count(*) / (

           select setting from pg_settings

           where name='shared_buffers')::integer,1)

        as buffer_percent,

        round(100.0*count(*)*8192 / pg_table_size(c.oid),1) as percent_of_relation

from pg_class c inner join pg_buffercache b on b.relfilenode = c.relfilenode inner

join pg_database d on ( b.reldatabase =d.oid and d.datname =current_database())

group by c.oid,c.relname order by 3 desc limit 100;




disk usage


select nspname,relname,pg_size_pretty(pg_relation_size(c.oid)) as "size"

from pg_class c left join pg_namespace n on ( n.oid=c.relnamespace)

where nspname not in ('pg_catalog','information_schema')

order by pg_relation_size(c.oid) desc limit 30;




March 22, 2018

PostgreSql create weekly child partition table

It seems that doing a quarterly partition is causing issues with the auto-vacum and clean up process, so instead now we are creating a single child table each week and remove the data every seven days,
so it took no time to figure it out and created the function trigger, but there should be a better way somewhere, but for now we still using inherit tables instead of actual file partitions.

replace FUNCTION weeklypartition_insert() 
returns TRIGGER language 'plpgsql' cost 100 volatile NOT leakproof AS $body$
DECLARE _partition_year VARCHAR(25);_partition              varchar(80);_IsNewChild             boolean;_current_week           varchar(4);_tablename              varchar(80);_currentdate timestamp ;BEGIN
  _currentdate := CURRENT_TIMESTAMP; 
  -- get current week 
  _current_week := date_part('week',new.insertdate)::text; ----- Week of the year (1-52) 
  -- get YYYY_MM format 
  _partition_year := to_char (new.insertdate,'YYYY')||'_'||date_part('week',new.insertdate); 
  _partition := tg_table_name || '_' || _partition_year; 
  _tablename := _partition; 
         SELECT 1 
         FROM   pg_tables 
         WHERE  schemaname= tg_table_schema 
         AND    tablename= _partition) THEN 
  raise notice 'A partition has been created %',tg_table_schema ||'.'|| _partition; 
  _isnewchild = true; 
  execute 'create table '|| tg_table_schema ||'.'|| _partition || ' (check( to_char (insertdate,''YYYY'||'_'|| _current_week||''') = '''|| to_char (new.insertdate,'YYYY')||'_'|| _current_week||''')) INHERITS ( ParentTableName);';
endIF;EXECUTE 'insert into '|| tg_table_schema || '.' || _partition || ' SELECT(' || tg_table_schema || '.' || tg_table_name || ' ' || quote_literal(new) || ').* RETURNING PrimaryKeyId;';IF (_isnewchild) then
-- Add primary key 
EXECUTE format 
  ('ALTER TABLE '||tg_table_schema||'.%s ADD PRIMARY KEY(PrimaryKeyId)', _partition); 
-- Add indexesEXECUTE format 
  ('CREATE INDEX ix_%s_PrimarykeyID ON '||tg_table_schema||'.%s USING btree ("SequenceID")', _tablename, _tablename);EXECUTE format
  ('CREATE INDEX ix_%s_Insertdate ON '||tg_table_schema||'.%s USING btree ("insertdate")', _tablename, _tablename);ENDIF;RETURN NULL;END;$BODY$;ALTER FUNCTION weeklypartition_insert() owner TO db_owner;
grant EXECUTE ON FUNCTION weeklypartition_insert() TO db_datawriter; 
grant EXECUTE ON FUNCTION weeklypartition_insert() TO PUBLIC; 
grant EXECUTE ON FUNCTION weeklypartition_insert() TO db_executor; 
grant EXECUTE ON FUNCTION weeklypartition_insert() TO db_owner;

Contact Form


Email *

Message *