Thursday, September 30, 2021

How to rename a database to use lower case instead of upper case characters

At my customer's site, we have standardized the Oracle SIDs (System ID) to use lower case character.

During an upgrade, the SID was accidently set to upper case. This new uppercase SID was picked up by Oracle Cloud Control and all of a sudden we had a non-standardized database within our ranks. Not a disaster, but irritating, nevertheless.

Given that your environment variable in you operating system is also set to lower case, you can simply restart the database with a new value for the parameter db_name in the parameter file to rectify the situation. There is no need to use the nid ("new id") tool or to recreate the controlfile. A simple restart with the right value for the parameter db_name will do.

It's easy to verify this conclution if you have access to another database which adheres to the standard. Check the value of the columnn name in v$database:
select name from v$database;

NAME
---------
PRODDB01

SQL> show parameter db_name

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
db_name                              string      PRODDB01

As expected, the query from v$database it is displayed in upper case as Oracle alwaysd does by default, but we also see that the value of the parameter db_name is actually in lower case.

In other words, there is no connection between the value of ORACLE_SID and the name of the database used in the controlfile. Oracle will always refer to the database in upper case. How we present and use this value in the operating system, is up to us.

These are the simple steps to get the value of the db_name parameter aligned with the value of your operating system variable ORACLE_SID:

Check your environment variable:
env | grep ORA
ORACLE_SID=testdb01 
It is indeed shown in lower case

Create a new pfile:
sqlplus / as sysdba
create pfile='inittestdb01.ora' from spfile;

Edit the pfile so that the db_name parameter has its value in lower case:
Old value:
*.db_name='TESTDB01'
New value:
*.db_name='testdb01'

Restart the database:
sqlplus / as sysdba
shutdown immediate
startup

If desirable, create a new spfile and restart the database one final time to instruct oracle to use it instead of the pfile:
create spfile from pfile='inittestdb01.ora';
shutdown immediate
startup

Tuesday, September 28, 2021

How to use opatch to check if Oracle Gateway for DRDA is installed

Log on to your oracle server, and type
cd $ORACLE_HOME/OPatch
./opatch lsinventory -oh /full/path/to/oracle_home -details > /home/oracle/inventory_details.log
The top lines of the generated file will reveal what's installed in your ORACLE_HOME:
cat /home/oracle/inventory_Details.log | more
Result:
Installed Top-level Products (2):

Oracle Database 12c                                                  12.1.0.2.0
Oracle Database Gateways                                             12.1.0.2.0
There are 2 products installed in this Oracle Home.
After that, more details about your installation is listed:

Installed Products (137):
Oracle Database Gateway for DRDA                                     12.1.0.2.0
Oracle Database Gateway for ODBC                                     12.1.0.2.0
Oracle Database Gateways                                             12.1.0.2.0

Wednesday, September 22, 2021

How to resolve ORA-48128 during export

Applicable to Oracle 19.3 and onwards.

The following error may occur when you export data using Data Pump

Parameter file:
userid=system/mypasswd
directory=DATA_PUMP_DIR
logfile=exp_sales.log
dumpfile=klm.dmp
exclude=statistics
job_name=exp_klm
schemas=KLM
flashback_time=systimestamp

Execute the export:
expdp parfile=expdp.par
Result:
RA-39126: Worker unexpected fatal error in KUPW$WORKER.CREATE_OBJECT_ROWS [USER]
USER:"GRDA_TEST"
ORA-39155: error expanding dump file name "/data1/export/klm.dmp"
ORA-48128: opening of a symbolic link is disallowed

ORA-06512: at "SYS.DBMS_SYS_ERROR", line 105
ORA-06512: at "SYS.KUPW$WORKER", line 12630
ORA-06512: at "SYS.DBMS_SYS_ERROR", line 105
ORA-06512: at "SYS.KUPF$FILE", line 9793
ORA-06512: at "SYS.KUPF$FILE_INT", line 1211
ORA-06512: at "SYS.KUPF$FILE", line 9734
ORA-06512: at "SYS.KUPW$WORKER", line 11341

----- PL/SQL Call Stack -----
  object      line  object
  handle    number  name
0xcf98d208     33543  package body SYS.KUPW$WORKER.WRITE_ERROR_INFORMATION
0xcf98d208     12651  package body SYS.KUPW$WORKER.DETERMINE_FATAL_ERROR
0xcf98d208     11612  package body SYS.KUPW$WORKER.CREATE_OBJECT_ROWS
0xcf98d208     15278  package body SYS.KUPW$WORKER.FETCH_XML_OBJECTS
0xcf98d208      3917  package body SYS.KUPW$WORKER.UNLOAD_METADATA
0xcf98d208     13746  package body SYS.KUPW$WORKER.DISPATCH_WORK_ITEMS
0xcf98d208      2439  package body SYS.KUPW$WORKER.MAIN
0xcf996200         2  anonymous block

KUPW: Object row index into parse items is: 1
KUPW: Parse item count is: 4
KUPW: In function CHECK_FOR_REMAP_NETWORK
KUPW: Nothing to remap
KUPW: In procedure BUILD_OBJECT_STRINGS - non-base info
KUPW: In procedure BUILD_SUBNAME_LIST with USER:KLM.KLM
KUPW: In function NEXT_PO_NUMBER
KUPW: PO number assigned: 13
KUPF$FILE.WRITE_LOB
KUPW: In procedure DETERMINE_FATAL_ERROR with ORA-39155: error expanding dump file name "/data1/export/klm.dmp"
ORA-06512: at "SYS.KUPF$FILE", line 9793
ORA-48128: opening of a symbolic link is disallowed
ORA-06512: at "SYS.KUPF$FILE_INT", line 1211
ORA-06512: at "SYS.KUPF$FILE", line 9734

Job "SYSTEM"."EXP_KLM" stopped due to fatal error at Wed Sep 22 08:42:49 2021 elapsed 0 00:00:04
Reason: bug

Solution:
Outlined in Doc ID 2654167.1 "Customer RecommendedDataPump Export (EXPDP) Fails Due to ORA-39155 ORA-48128":
ALTER SYSTEM  SET "_disable_directory_link_check" =true  SCOPE=SPFILE;
shutdown immediate
startup

Monday, September 20, 2021

PostgreSQL: how to use the to_date, to_char and cast conversion functions

Building on the example in my previous post, this is how you can use the to_date and to_char functions for predicates used against both the original, non partitioned table, and the partitioned version of the same.

Query the non-partitioned table, you could first convert to character, and use the desired format. After that you can convert to date:
select *
from myuser.mytable
where to_date(to_char(created_date,'YYYY.MM.DD'),'YYYY.MM.DD') between '2020.01.01' and  '2020.01.31';

To query the partitioned table, you need an an explicit cast to the same data type used by the partitioned key column:
select *
from myuser.mytable
where  cast(created_date as date) between '2020.01.01' and  '2020.01.31';
I realized this difference when the number of rows returned was not what I expected. After I changed my SQL statement to use CAST, the number of rows was exactly what I expected.

PostgreSQL: how to create a partitioned table using a conversion function

Consider the following table:
CREATE TABLE myuser.mytable (
    id integer NOT NULL,
    created_date timestamp without time zone NOT NULL,
    event_json text NOT NULL,
    format_version text NOT NULL,
    rep_id text NOT NULL,
    appversion text NOT NULL
);
Let's say you want to partition this table by RANGE, and you would like to use the column "created_date" which is of data type "TIMESTAMP WITHOUT TIME ZONE".
You could use a conversion function like CAST to convert the column to a partitioned one, like this:
  CREATE TABLE myuser.mytable (
    id integer NOT NULL,
    created_date timestamp without time zone NOT NULL,
    event_json text NOT NULL,
    format_version text NOT NULL,
    rep_id text NOT NULL,
    appversion text NOT NULL
)
partition by range(cast(created_date as date));
Obviously you proceed with adding your partitions the way you normally would in PostgreSQL, for example:
CREATE TABLE myuser.mytable_p_202001
 PARTITION OF myuser.mytable
 FOR VALUES FROM ('2020.01.01') TO ('2020.01.31');

CREATE TABLE myuser.mytable_p_202002
 PARTITION OF myuser.mytable
 FOR VALUES FROM ('2020.02.01') TO ('2020.02.29');

etc

CREATE TABLE myuser.mytable_p_default
 PARTITION OF myuser.mytable
 DEFAULT;

Beware: If you try to create a primary key constraint on the partition key column in this case, you will receive
DETAIL:  PRIMARY KEY constraints cannot be used when partition keys include expressions.

You are of course free to add a normal search index on the column:
CREATE INDEX myidx1 ON myuser.mytable USING btree ( cast (created_date as date) );
Read more about the CAST function in the documentation
Another good source is postgressqltutorial.com

Friday, September 17, 2021

How to solve error in pg_top stating "relation does not exist"

When you execute pg_top, you get to see the top database sessions in a way familiar to most people with experience from Unix environments.



There are many options to pg_top, and one that I wanted to use, was
E       - show execution plan (UPDATE/DELETE safe)
When you press E, you are prompted for a process ID:
Re-determine execution plan: 119692

But I kept getting the message:
Query Plan:

ERROR:  relation "myschema.mytable" does not exist
LINE 4:     from myschema.mytable
                 ^
Hit any key to continue:

Solution: Execute pg_top with database name, host name, port, username and password, like this:
$ pg_top -d mydb -h myhost.mydomain.com -p 5432  -U mysuperuser -W
After I executed pg_top in this manner, the E option worked as exected and displayed the execution plan.

Thursday, September 16, 2021

How to list the contents of a custom format PostgreSQL export file

List the content of a "custom format" export file created by pg_dump by specifying the -l flag:

pg_restore /data1/exports/musicdb_Fc.dmp -l


List the content of a "directory format" export file created by pg_dump by specifying the -l flag. Note that the path should point to the directory in which the toc.dat file resides, in my case /data1/exports/mydir:
pg_restore /data1/exports/mydir/ -l


In both cases, output will be something akin to
;
; Archive created at 2021-09-16 12:51:17 CEST
;     dbname: musicdb
;     TOC Entries: 29
;     Compression: -1
;     Dump Version: 1.13-0
;     Format: CUSTOM
;     Integer: 4 bytes
;     Offset: 8 bytes
;     Dumped from database version: 11.12
;     Dumped by pg_dump version: 11.12
;
;
; Selected TOC Entries:
;
10; 2615 45112854 SCHEMA - music postgres
3764; 0 0 ACL - SCHEMA music postgres
202; 1259 45112865 TABLE music albums music
203; 1259 45112868 TABLE music albums_1974 music
204; 1259 45112871 TABLE music albums_1979 music
201; 1259 45112863 SEQUENCE music albums_album_id_seq music
205; 1259 45112874 TABLE music albums_default music
207; 1259 45112879 TABLE music artists music
206; 1259 45112877 SEQUENCE music artists_artist_id_seq music
209; 1259 45112884 TABLE music bands music
208; 1259 45112882 SEQUENCE music bands_band_id_seq music
200; 1259 45112858 TABLE music songs music
199; 1259 45112856 SEQUENCE music songs_song_id_seq music
3750; 0 45112868 TABLE DATA music albums_1974 music
3751; 0 45112871 TABLE DATA music albums_1979 music
3752; 0 45112874 TABLE DATA music albums_default music
3754; 0 45112879 TABLE DATA music artists music
3756; 0 45112884 TABLE DATA music bands music
3748; 0 45112858 TABLE DATA music songs music
3765; 0 0 SEQUENCE SET music albums_album_id_seq music
3766; 0 0 SEQUENCE SET music artists_artist_id_seq music
3767; 0 0 SEQUENCE SET music bands_band_id_seq music
3768; 0 0 SEQUENCE SET music songs_song_id_seq music
3620; 2606 45112862 CONSTRAINT music songs songs_pkey music