Re: [GENERAL] [PERFORM] Very slow query in PostgreSQL 9.3.3

2014-03-13 Thread Ian Lawrence Barwick
2014-03-14 4:26 GMT+09:00 : > PostgreSQL 9.3.3 RHEL 6.4 > > Total db Server memory 64GB (...) > effective_cache_size = 48MB I'm not sure if this will help directly, but is the value for 'effective_cache_size' intentional? 48 *GB* would be a more likely setting. Regards Ian Barwick -- Sent v

Re: [GENERAL] [BUGS] Very slow query in PostgreSQL 9.3.3

2014-03-13 Thread Sergey Konoplev
On Thu, Mar 13, 2014 at 12:26 PM, wrote: > *** Problem Query *** > > explain (analyze on, buffers on) Select * from measurement this_ > where this_.logdate between '2007-12-19 > 23:38:41.22'::timestamp and '2007-12-20 08:01:04.22'::timestamp >

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Susan Cassidy
The only one that comes out different is inet_server_addr, via the program, it comes out: ::1 whereas via psql it comes out empty. Yes, I am 100% sure I am using the same schema (which I never specify, so I am using 'public') and the same user and database. Susan On Thu, Mar 13, 2014 at 1:53

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Susan Cassidy
I finally figured it out, after changing my code to output the lines per the number of rows of output, instead of until data[0] was blank. It turned out that data[0] was sometimes blank, and I forgot about that, and was stopping the output after I got back an empty record (or so I thought). So, a

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Rob Sargent
On 03/13/2014 03:23 PM, Susan Cassidy wrote: The only one that comes out different is inet_server_addr, via the program, it comes out: ::1 whereas via psql it comes out empty. Yes, I am 100% sure I am using the same schema (which I never specify, so I am using 'public') and the same user and

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Susan Cassidy
You asked for it (by the way, just p is asking for a lot of output): root23 0.0 0.0 0 0 ?SFeb20 0:00 [cgroup] root24 0.0 0.0 0 0 ?SFeb20 0:00 [khelper] root27 0.0 0.0 0 0 ?SFeb20 0:00 [pm] root

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Rob Sargent
On 03/13/2014 02:44 PM, Susan Cassidy wrote: No, I don't have 2 instances running. I default the port on the psql command line, and the perl program is using 5432, as normal. Now, I'm discovering that syslog is no longer logging anything. I bounced it, but to no avail. Susan On Thu, Mar

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Rodrigo Gonzalez
On Thu, 13 Mar 2014 13:44:48 -0700 Susan Cassidy wrote: > No, I don't have 2 instances running. I default the port on the psql > command line, and the perl program is using 5432, as normal. > > Now, I'm discovering that syslog is no longer logging anything. I > bounced it, but to no avail. Th

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Dennis Jenkins
1) Run both "psql" and "perl" under "strace" and search the output for which sockets it connects to. eg, strace -o /tmp/psql.log psql -Upgsql -dmydatabase -c"select version();" 2) Add a query into your perl script to perform the following SQL and print the results: select current_database();

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Susan Cassidy
No, I don't have 2 instances running. I default the port on the psql command line, and the perl program is using 5432, as normal. Now, I'm discovering that syslog is no longer logging anything. I bounced it, but to no avail. Susan On Thu, Mar 13, 2014 at 1:34 PM, Rodrigo Gonzalez wrote: > O

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Rodrigo Gonzalez
On Thu, 13 Mar 2014 13:28:38 -0700 Susan Cassidy wrote: > No, it is connecting to localhost, which is the same system I am > running psql on. > > Susan > Well, if one query is logged and the other one is not it means that it is running against different servers (as far as I understand logging).

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Susan Cassidy
No, it is connecting to localhost, which is the same system I am running psql on. Susan On Thu, Mar 13, 2014 at 1:26 PM, Rodrigo Gonzalez wrote: > On Thu, 13 Mar 2014 13:20:53 -0700 > Susan Cassidy wrote: > > > Another odd thing is that it is apparently not logging statements from > > Perl, o

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Rodrigo Gonzalez
On Thu, 13 Mar 2014 13:20:53 -0700 Susan Cassidy wrote: > Another odd thing is that it is apparently not logging statements from > Perl, only from psql. I don't know why. I thought I had it set up > right to log to syslog. I've had good luck before with that on other > installations. > I can

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Steve Atkins
On Mar 13, 2014, at 1:20 PM, Susan Cassidy wrote: > Yes, I am running with use strict. The statement I pasted in is after perl > quoting, being written out by the same perl program. I just take that > statement and paste it into the psql window. > > DBI->trace showed nothing out of the ord

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Susan Cassidy
No, I am connecting to the right database, as my log info shows. No duplicate table names, except in different databases. Those tables show all the data I expect, oddly enough, if I connect the perl program to that database. It is only this database that is giving me trouble. Which is another o

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Susan Cassidy
Yes, I am running with use strict. The statement I pasted in is after perl quoting, being written out by the same perl program. I just take that statement and paste it into the psql window. DBI->trace showed nothing out of the ordinary. It just shows the lines being fetched that I am seeing in

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Steve Atkins
On Mar 13, 2014, at 12:18 PM, Susan Cassidy wrote: > I have a weird issue that I can't figure out. > > If I run the exact same query through psql as through perl DBI, I get > different results. I get far fewer results with DBI than through the psql > command line. > > Any ideas why that wo

Re: [GENERAL] [BUGS] Very slow query in PostgreSQL 9.3.3

2014-03-13 Thread Pavel Stehule
2014-03-13 20:26 GMT+01:00 : > PostgreSQL 9.3.3 RHEL 6.4 > > Total db Server memory 64GB > > > # - > # PostgreSQL configuration file > # - > max_connections = 100 > shared_buffers = 16GB > work_mem = 32MB > maintenance_work_mem = 1GB > seq_pa

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread john gale
On Mar 13, 2014, at 12:18 PM, Susan Cassidy wrote: > I copied and pasted the query from the program's log file, so I know I'm > doing the exact same query. If it matters, I'm only seeing the rows with > 'root' in them via DBI, which the CASE statement refers to. Try enabling full query l

[GENERAL] Very slow query in PostgreSQL 9.3.3

2014-03-13 Thread fburgess
PostgreSQL 9.3.3 RHEL 6.4Total db Server memory 64GB# -# PostgreSQL configuration file# -max_connections = 100shared_buffers = 16GBwork_mem = 32MB maintenance_work_mem = 1GBseq_page_cost = 1.0   random_p

Re: [GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Tom Lane
Susan Cassidy writes: > I have a weird issue that I can't figure out. > If I run the exact same query through psql as through perl DBI, I get > different results. I get far fewer results with DBI than through the psql > command line. Any possibility that the perl program is connecting to a diffe

Re: [GENERAL] can't cast hstore to json

2014-03-13 Thread Joe Van Dyk
On Thu, Mar 13, 2014 at 12:15 PM, Joe Van Dyk wrote: > I've got a postgresql 9.3.2 server, compiled from scratch. Getting this > error: > > # select hstore('a', 'b')::json; > ERROR: cannot cast type hstore to json > LINE 1: select hstore('a', 'b')::json; >^ > # se

Re: [GENERAL] can't cast hstore to json

2014-03-13 Thread Tom Lane
Joe Van Dyk writes: > I've got a postgresql 9.3.2 server, compiled from scratch. Getting this > error: > # select hstore('a', 'b')::json; > ERROR: cannot cast type hstore to json Probably need to update the hstore extension to 1.2. regards, tom lane -- Sent via pgsql

[GENERAL] puzzling perl DBI vs psql problem

2014-03-13 Thread Susan Cassidy
I have a weird issue that I can't figure out. If I run the exact same query through psql as through perl DBI, I get different results. I get far fewer results with DBI than through the psql command line. Any ideas why that would be? The query is: SELECT st.description, st.scene_thing_instan

[GENERAL] can't cast hstore to json

2014-03-13 Thread Joe Van Dyk
I've got a postgresql 9.3.2 server, compiled from scratch. Getting this error: # select hstore('a', 'b')::json; ERROR: cannot cast type hstore to json LINE 1: select hstore('a', 'b')::json; ^ # select array_to_json(array[hstore('a', 'b')]); array_to_json -

Re: [GENERAL] FATAL: the database system is starting up

2014-03-13 Thread fluxh
I have the same problem. I have a log like patrick keshishian. I have not a backup. I don't know that do. Help me please!!! Information: - Server with RAID1 on SAS HD Hot-plug. - Ubuntu Server 12.04 x64. - Postgresql 9.1. When I do "psql" PostgreSQL shows "psql: FATAL: the database system is s

Re: [GENERAL] UNIQUE contraint that is initially deferred in Postgres 8.4.13

2014-03-13 Thread Johannes Bauer
On 13.03.2014 13:02, Magnus Hagander wrote: > My guess is you're looking in the wrong place in the docs - can you specify > where you are looking? If you are in the right place then the docs are > wrong. Here: http://www.postgresql.org/docs/8.4/static/sql-createtable.html > Deferrable unique con

Re: [GENERAL] UNIQUE contraint that is initially deferred in Postgres 8.4.13

2014-03-13 Thread Magnus Hagander
On Thu, Mar 13, 2014 at 12:59 PM, Johannes Bauer wrote: > Hi list, > > I'm having trouble with a UNIQUE constraint that I need to have > DEFERRABLE INITIALLY DEFERRED. On my Dev machine (Postgres 9.1) it works > fine: > > alter table foo drop constraint bar; > ALTER TABLE > > alter table foo add c

[GENERAL] UNIQUE contraint that is initially deferred in Postgres 8.4.13

2014-03-13 Thread Johannes Bauer
Hi list, I'm having trouble with a UNIQUE constraint that I need to have DEFERRABLE INITIALLY DEFERRED. On my Dev machine (Postgres 9.1) it works fine: alter table foo drop constraint bar; ALTER TABLE alter table foo add constraint bar UNIQUE (col1, col2) deferrable initially deferred; ALTER TAB