| Author |
Message |
|
|
|
Still doesn't work
|
 |
|
|
No, it is not
As I said above:
The only way to move from one platform / architecture to another is to use a 'logical' method:
Meaning that either the OS or the platform (32 / 64 bit... edian-ness etc...) must be the same.
|
 |
|
|
No, it isn't. Physical (PITR) backups are based on copying the physical files directly off the database server. The only way to move from one platform / architecture to another is to use a 'logical' method:
pg_dump / restore
pg_dump / psql
slony replication
londiste replication (untested)
bucardo replication (untested)
database links
Or something like that, you cannot port the physical datafiles of postgres from one platform / architecture to another.
|
 |
|
|
is there a plugin to postgres studio available... etc ?
If you're just looking to compare two schemas, you can do a 'pg_dump -s databasename > file.sql ' and run a 'diff' on it.
If you're looking for a decent data-modeler that directly supports postgres, take a look at 'power architect' (free: http://www.sqlpower.ca/page/architect)
|
 |
|
|
OHH!!! I'm sorry! You're trying to go FROM sql server to postgres?
I'm not [by any means] a sql server DBA, but I do seem to remember that there was a change from sql server 2005 -> 2008 regarding what type of data sources you could connect to. But that may have just been in regards to SSIS. If you can make the ODBC connection to postgres, you should be okay. On unix, you could use isql to test that the odbc source was setup correctly, on windows, I'm not sure what the ODBC test program is ( if there is one ).
Take a look at articles like:
http://www.developer.com/open/article.php/631251/Connecting-to-PostgreSQL-from-Windows-Platforms.htm
for more information.
Sorry I got confused!
|
 |
|
|
Gotcha,
Well, there's nothing 'standard' in the postgres database to do this... BUT, I have done work in the past with dblink_tds and postgres where I had a direct link into SQL server. It was a bit 'leaky' as far as memory goes, but it was fast and worked really well:
http://pgfoundry.org/projects/dblink-tds/
|
 |
|
|
Hey Joe,
Are you trying to link the database to engine to sql server.... or just a one time data migration?
|
 |
|
|
You should set your PATH to /Library/PostgreSQL/8.4/bin
DYLD_LIBRARY_PATH=/Library/PostgreSQL/8.4/lib
and before running configure
export CFLAGS="-L/Library/PostgreSQL/8.4/lib"
I don't know anything (read: AT ALL about ruby builds, but it just looks like the ruby configure scripts aren't finding the library folder.
|
 |
|
|
You can't get a recompiled EnterpriseDB... but you really shouldn't recompile the postgresql core just for the blocksize. My guess is that you're using too many columns of fixed-sized data in your table (600 is a lot) Is there any way to normalize that table?
|
 |
|
|
13/03/2010 07:53:29 Subject: Lost postgres password
Hi
I can`t setup postgresql-8.4.2-1windows because password is inccorect and i don`t know password...
please help
First -- You should start a new topic for this... people will find your question more easily that way.
Second, try going into the windows user manager and resetting the 'postgres' user's password from there. Then, when you run the installer enter that password
|
 |
|
|
How did your database crash, what was the log message?
If your system crashed, just move the whole c:\data directory to the new machine, be sure that both machines are running the same OS and are the same architecture.
|
 |
|
|
Read the documentation
http://www.enterprisedb.com/documentation/app-createuser.html
|
 |
|
|
surely there must be a method to deploy Postgres without an account being created?
Due to the architecture of postgres, there is not a way to do this from the installer. You could however script the action. Using something like sc.exe (http://support.microsoft.com/kb/251192)
|
 |
|
|
Try using a directory other than the default for data... try 'c:\pgdata'
Does it help?
|
 |
|
|
It looks like a permissions issue. Try setting the data directory to something other than the default. For example, try using
c:\pgdata
And see what happens.
|
 |
|
|
Sorry for the delay Kiran, I've been traveling a lot lately.
I think the problem you're having is the double quotes surrounding public and the '.' before the word database. I think the syntax you want is:
CREATE PUBLIC DATABASE LINK ....
Take a look at the docs:
http://www.enterprisedb.com/docs/en/8.3/oracompat/EnterpriseDB_OraCompat_EN_8.3-49.htm
|
 |
|
|
|
I seem to recall that in the release 2 version of 8.3, this is a limitation.
|
 |
|
|
We've seen this issue with automounting homedirs before and the new installers will be going through this test cycle for the 8.4 release (when it comes out). In the meantime, I'll see if I can get a developer to post the manual instructions on installing the 'edbnetwork' directory up here so that you can fall back on the 'extractall'
|
 |
|
|
If you look in your dbserver_logs or pg_logs directory, do you see the SQL statement that caused the error?
|
 |
|
|
This is for a couple of reasons:
*) rownum is an operation on the result set, not the actual data
*) B/c of the above, there are some things to note:
- If you don't require a sort, if each database returns things slightly differently (due to MVCC) then the order will be different and rownums will be out of sync
- In your architecture pgPool-II is responsible for returning the results to the client, since it has no concept of the 'rownum' variable, it will not be able to return proper data to the system
The net? Since pgPool-II has not been modified to include the Oracle language extensions, rownum will probably not work through it. You would have to either modify pgPool-II or directly query a database to use the rownum. I do however recommend that you use the LIMIT and OFFSET keywords as they are natively supported in all systems.
|
 |
|
|
What are your log settings? Do you have anything set to 'debugx' level?
|
 |
|
|
You would set this in the 'postgresql.conf' file of your Advanced Server instance. The configuration entry doesn't exist in the default file, just simply edit the file and add:
anywhere within the file.
|
 |
|
|
|
OKay, but you still haven't done the test I suggested. Until you test and VERIFY that the port is open, you don't know that it's working.
|
 |
|
|
Upgrade to 8.3R2 and use the Oracle style dblinks:
http://www.enterprisedb.com/docs/en/8.3/oracompat/EnterpriseDB_OraCompat_EN_8.3-49.htm
You'll need to be running Advanced Server 8.3R2 however. If you want to use postgres, use postgres dblinks:
http://www.postgresql.org/docs/current/static/dblink.html
--Scottie
|
 |
|
|
Raghavendra_R wrote:Hi Scott,
Thank you..
No problem
Is there any particular reason why it is not accepted in VALUES clause..
Yeah, 'generate_series()' is a special type of function called a Set-Returning function (SRF). These types of functions return a set of items, not just one item. In actuality, an SRF is basically the same thing as a table. The problem with SRF's is that they need to be referenced as if they are a table or there is an error.
--Scottie
|
 |
|
|