Re: [ADMIN] Log entry: srm: Permission Denied (Answered)
Sorry, developer was checking DELETE permissions for table srm. On Wednesday 22 May 2002 04:31 pm, William Meloney wrote: > Dear listers: > > The following appeared in my Postgresql.log file > > ERROR: srm: Permission Denied > > Any thoughts or observations? > > Thank you. > > Best regards. > > William Meloney > bmeloney@mindspringcom > 2002-05-22 > > > > ---(end of broadcast)--- > TIP 4: Don't 'kill -9' the postmaster ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
[ADMIN] Log entry: srm: Permission Denied
Dear listers: The following appeared in my Postgresql.log file ERROR: srm: Permission Denied Any thoughts or observations? Thank you. Best regards. William Meloney bmeloney@mindspringcom 2002-05-22 ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [ADMIN] Performance Problem Index Ignored, but why
I'm not sure how well this works in Postgres. For Informix, we could trick the optimizer into using an index with something like: Select colname from table where colname !=NULL. Specifically mentioning the column in the query was the trick. In esql/C, there were return parameters that then told you how many rows were found. >CREATE INDEX st_v_state_idx ON state_tst USING btree (v_state); >CREATE INDEX st_f_state_idx ON state_tst USING btree (f_state); > >Load the table using a copy from ... > >vacuum verbose analyze state_tst; > >Total rows: 14309241 > >Queries using either f_state = or v_state = explain (and appear to >execute) using a sequential scan. Resulting in 60 - 80 second query >times. > >Can I force the use of an index? Or do I have something wrong? Any >ideas? > >pg_test=# explain select count(*) from state_tst where f_state = 'PA'; >NOTICE: QUERY PLAN: > >Aggregate (cost=277899.65..277899.65 rows=1 width=0) > -> Seq Scan on state_tst (cost=0.00..277550.51 rows=139654 > width=0) > >EXPLAIN > >---(end of broadcast)--- >TIP 4: Don't 'kill -9' the postmaster ---(end of broadcast)--- TIP 2: you can get off all lists at once with the unregister command (send "unregister YourEmailAddressHere" to [EMAIL PROTECTED])
Re: [ADMIN] Performance Problem Index Ignored, but why
> estimate that 139654 rows will match f_state = 'PA' in the right No, 375342 is the actual number. Using the index does appear slower (limited testing noted). explain select count(*) from state_tst where f_state = 'PA'/ NOTICE: QUERY PLAN: Aggregate (cost=277899.65..277899.65 rows=1 width=0) -> Seq Scan on state_tst (cost=0.00..277550.51 rows=139654 width=0) select count(*) from state_tst where f_state = 'PA'/ count '375342' Elapsed: 139 wallclock secs set enable_seqscan = off/ explain select count(*) from state_tst where f_state = 'PA'/ NOTICE: QUERY PLAN: Aggregate (cost=542303.53..542303.53 rows=1 width=0) -> Index Scan using st_f_state_idx on state_tst (cost=0.00..541954.39 rows=139654 width=0) select count(*) from state_tst where f_state = 'PA'/ count '375342' Elapsed: 222 wallclock secs Tom On Wed, May 22, 2002 at 12:26:35AM -0400, Tom Lane wrote: > "Thomas A. Lowery" <[EMAIL PROTECTED]> writes: > > Can I force the use of an index? > > Try "set enable_seqscan = off". But on the basis of what you've shown, > it's not obvious that an indexscan will be faster. Is the planner's > estimate that 139654 rows will match f_state = 'PA' in the right > ballpark? > > regards, tom lane ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/users-lounge/docs/faq.html
[ADMIN] Problem in 'User Securities' in postgres
Hi All, I gave my previous mail on 'user securities' with subject 'probelm in usersecurities'. In that i mentioned that access to database on localhost is allowed without password. I got suggestion that to go through 'clinet authentication'. As per that document i changed the 'pg_hba.conf' file as follows. local all password while accessing database it is asking for password, if i give database password, i am getting error as 'password authentication failure for user '' ' . host all x.x.x.x x.x.x.x password If i try to connect through pg_access (tcp/Ip) connection. If I supplied user password i got same above error. In case of 'trust' connection it is allowing every body without password checking. How can i use password authentication. Another question shall we use 'unix securities' using Postgres like 'windows-nt and sqlserver'. and what security protocol is maintaining by postgres. thanks alot for any suggestion. with best regards, bhaskar
[ADMIN] Performance Problem Index Ignored, but why
I've the task of porting a current Oracle application to PostgreSQL. Database: 7.2.1 OS: Linux 2.4.9-13smp I've an odd thing happening with a query. Using a simple table: Table "state_tst" Column | Type | Modifiers -+--+--- id | integer | not null v_state | character varying(2) | f_state | character(2) | Indexes: st_f_state_idx, st_v_state_idx Primary key: state_tst_pkey id is a sequence number and primary key, v_state and f_state are 2 character U.S. States. I created v_state as varchar(2) and f_state as char(2) to test if the query explained/performed differently (it doesn't). CREATE INDEX st_v_state_idx ON state_tst USING btree (v_state); CREATE INDEX st_f_state_idx ON state_tst USING btree (f_state); Load the table using a copy from ... vacuum verbose analyze state_tst; Total rows: 14309241 Queries using either f_state = or v_state = explain (and appear to execute) using a sequential scan. Resulting in 60 - 80 second query times. Can I force the use of an index? Or do I have something wrong? Any ideas? pg_test=# explain select count(*) from state_tst where f_state = 'PA'; NOTICE: QUERY PLAN: Aggregate (cost=277899.65..277899.65 rows=1 width=0) -> Seq Scan on state_tst (cost=0.00..277550.51 rows=139654 width=0) EXPLAIN ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
[ADMIN] Rserv and Boolean...
Fogive me if this is the wrong forum for this question, but I have a bit of an issue with rserv and Postgres 7.1.3. The database that we have has several boolean datatypes, but when I try to either prepare a snap shot, or run the Replicate command (which seems to prepare a snap shot anyway), I get the following error in my postgres logs. ERROR: Cannot cast type 'bool' to 'text' Does anyone know of a way around this other that changing data types in the DB? Thanks in advance! -James ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org