pgAdminIII performance issue - Mailing list pgadmin-support
Just installed pgAdmin 1.22.0 on a new desktop when when I noticed some basic queries started running extremely slowly. I went back and installed 1.18.1 (which was being used on my previous desktop) and ran them concurrently to better attempt to understand what is going on. The following describes the issue:
pgAdmin III v1.22.0 (standard install/no build) running Windows 7, 5 gen i5/8GB ram, Gigabit networking.
Here are the results of running of two versions of pgAdmin running on the same windows desktop against the same backend server; The two have very different outcomes:
Version | v1.18.1 | V1.22.0 |
Connection | Same connection (host, database name, port, username, and query) using the same windows desktop. | |
Query | select * from m.events where time > '2016-01-01' | |
Execution Time | ~1.8 seconds Execution time min ~1.6s max 1.9s over 30 executions. | ~25 seconds Execution time min ~25.5 max 25.9s over 30 executions. |
Explain | "Append (cost=0.43..5122.55 rows=3920 width=200)" " -> Index Scan using events_time_idx on events (cost=0.43..5077.55 rows=3521 width=203)" " Index Cond: (time > '2016-01-01 00:00:00-06'::timestamp with time zone)" " -> Seq Scan on events_2002 (cost=0.00..15.00 rows=133 width=172)" " Filter: (time > '2016-01-01 00:00:00-06'::timestamp with time zone)" " -> Seq Scan on events_2003 (cost=0.00..15.00 rows=133 width=172)" " Filter: (time > '2016-01-01 00:00:00-06'::timestamp with time zone)" " -> Seq Scan on events_2004 (cost=0.00..15.00 rows=133 width=172)" " Filter: (time > '2016-01-01 00:00:00-06'::timestamp with time zone)" | "Append (cost=0.43..5122.55 rows=3920 width=200)" " -> Index Scan using events_time_idx on events (cost=0.43..5077.55 rows=3521 width=203)" " Index Cond: (time > '2016-01-01 00:00:00-06'::timestamp with time zone)" " -> Seq Scan on events_2002 (cost=0.00..15.00 rows=133 width=172)" " Filter: (time > '2016-01-01 00:00:00-06'::timestamp with time zone)" " -> Seq Scan on events_2003 (cost=0.00..15.00 rows=133 width=172)" " Filter: (time > '2016-01-01 00:00:00-06'::timestamp with time zone)" " -> Seq Scan on events_2004 (cost=0.00..15.00 rows=133 width=172)" " Filter: (time > '2016-01-01 00:00:00-06'::timestamp with time zone)" |
Additional | · I ran test concurrently (same query run at exact same time using the two different versions of pgAdmin on the same windows 7 desktop, against the same dataset and server) to what would happen:
· Perform a count “select count(*) where time > '2016-01-01'” both returned in 12 milliseconds.
· Events table has ~14,000,000 rows. | |
|
| |
We’re using PostgresSQL 9.3 (eventually moving to 9.4/9.5).
We also know that we’re returning a lot of data but there are reasons for this.
(PNG file of table listed above)
Has anyone else seen this or can anyone reproduce similar results?
Contact me for additional information.
-Derek
pgadmin-support by date: