Friday, April 13, 2018

A solution for ORA-28040



When connecting to Oracle 12c databases you may get

ERROR:
ORA-28040: No matching authentication protocol

upon login.

The error is defined as follows:

28040, 0000, "No matching authentication protocol"
Cause: There was no acceptable authentication protocol for either client or server.

Action: The administrator should set the values of the SQLNET.ALLOWED_LOGON_VERSION_SERVER and SQLNET.ALLOWED_LOGON_VERSION_CLIENT parameters, on both the client and on the server, to values that match the minimum version software supported in the system. This error is also raised when the client is authenticating to a user account which was created without a verifier suitable for the client software version. In this situation, that account's password must be reset, in order for the required verifier to be generated and allow authentication to proceed successfully.


Solution: edit your $TNS_ADMIN/sqlnet.ora file to include this directive:

SQLNET.ALLOWED_LOGON_VERSION_SERVER=8

which will allow older clients to connect. Oracle claims that the default value after an upgrade to Oracle 12.1 is 11, but I have found that the parameter needs to be explicitly set in the server's sqlnet.ora file before older clients are allowed.

Note that the parameter SQLNET.ALLOWED_LOGON_VERSION_CLIENT is relevant only when the database server itself acts like a client against other database servers. Consequeltly, unless I know for certain that a database server is actually used as a client against other remote database servers, I will specify them like this for my Oracle 12c database servers:

SQLNET.ALLOWED_LOGON_VERSION_SERVER=11
SQLNET.ALLOWED_LOGON_VERSION_CLIENT=12


Check the Oracle documentation for more info

Tuesday, April 10, 2018

How to change the start time for a maintenance window

For example, if you want to change the start time for the MONDAY_WINDOW from the default 22:00 to 23:59, use the following syntax:

BEGIN
DBMS_SCHEDULER.DISABLE(name=>'"SYS"."MONDAY_WINDOW"',force=>TRUE);
END;
/

BEGIN
DBMS_SCHEDULER.SET_ATTRIBUTE(name=>'"SYS"."MONDAY_WINDOW"',attribute=>'REPEAT_INTERVAL',value=>'FREQ=WEEKLY;BYDAY=MON;BYHOUR=23;BYMINUTE=59;BYSECOND=0');
END;
/

BEGIN
DBMS_SCHEDULER.ENABLE(name=>'"SYS"."MONDAY_WINDOW"');
END;
/

More on finding failed autostats jobs



In a previous post, I showed how to check if an autostats job had succeeded.

Recently I had to unlock table stats on a very large partitioned table, and then lock down older partition stats explicitly, to allow Oracle to gather statistics for the most recent partitions automatically with the auto stats job.

I wanted to know if Oracle was able to analyze the partitions within the designated 4 hours of weekday window.
When you read the code snippet below, please keep in mind that I have defined my weekday windows to start at 23:59 every evening.

I found the following query to be useful:

select distinct 
       opid, 
       target,
       target_type, 
       to_char(start_Time,'dd.mm.yyyy hh24:mi:ss') "starttime",
       to_char(end_time,'dd.mm.yyyy hh24:mi:ss')   "endtime",
       status
from dba_optstat_operation_tasks 
where opid IN (
            SELECT ID
            FROM DBA_OPTSTAT_OPERATIONS
            WHERE OPERATION LIKE 'gather_database_stats (auto)'
            AND TO_CHAR(START_TIME,'dd.mm.yyyy hh24:mi') = '09.04.2018 23:59'
)
and target like '%SALES%';
The result:

OPID TARGET TARGET_TYPE starttime endtime STATUS
47815
SH.SALES TABLE 09.04.2018 23:59:39 10.04.2018 03:25:17 COMPLETED
47815
SH.SALES.P_SALES_DATA_2016 TABLE PARTITION 10.04.2018 03:25:41 10.04.2018 03:59:00 TIMED OUT
47815
SH.SALES.P_SALES_DATA_MAX TABLE PARTITION 10.04.2018 03:25:33 10.04.2018 03:25:33 COMPLETED

All in all, the autostat job started on April 9th, 2018 at 23:59, had the following outcome:
select status,count(*)
from dba_optstat_operation_tasks 
where opid IN (
            SELECT ID
            FROM DBA_OPTSTAT_OPERATIONS
            WHERE OPERATION LIKE 'gather_database_stats (auto)'
            AND TO_CHAR(START_TIME,'dd.mm.yyyy hh24:mi') = '09.04.2018 23:59'
)
group by status;

STATUS COUNT(*)
FAILED
1
TIMED OUT
12
COMPLETED
339

From the output you can now proceed to find out why the analysis failed or timed out. The query below will list these objects for you:

select distinct opid, 
       target,
       target_type, 
    to_char(start_Time,'dd.mm.yyyy hh24:mi:ss') "starttime",
    to_char(end_time,'dd.mm.yyyy hh24:mi:ss')   "endtime",
    status
from dba_optstat_operation_tasks 
where opid IN (
            SELECT ID
            FROM DBA_OPTSTAT_OPERATIONS
            WHERE OPERATION LIKE 'gather_database_stats (auto)'
            AND TO_CHAR(START_TIME,'dd.mm.yyyy hh24:mi') = '09.04.2018 23:59'
)
and STATUS IN ('TIMED OUT','FAILED');

Thursday, March 15, 2018

The error

SP2-0552: Bind variable "B2" not declared.

can arise when using incorrect declarations in your SQL scripts.

For example, I had the follwing in a script:
var B2 DATE;
EXEC :B2 := to_date('22.02.2018','dd.mm.yyyy');

For sqlplus to accept this string, you need to declare your string as VARCHAR2, even though you intend to use datatype DATE in your query.

Declare it as VARCHAR instead:

var B2 VARCHAR2(10);


How to load a plan from the cursor cache into the SMB



set serveroutput on
VARIABLE cnt NUMBER

-- Specifying both the SQL ID and the plan hash value:
EXECUTE :cnt := DBMS_SPM.LOAD_PLANS_FROM_CURSOR_CACHE( -
                    sql_id => '5abzqhtfcvr73' -
                    ,plan_hash_value =>7104589 -
                    ,fixed => 'YES' -
                    ,enabled=>'YES');


-- without a specific plan hash value:
EXECUTE :cnt := DBMS_SPM.LOAD_PLANS_FROM_CURSOR_CACHE( -
                    sql_id => '5abzqhtfcvr73' -
                    ,fixed => 'YES' -
                    ,enabled=>'YES');
print :cnt;
exit

Wednesday, March 14, 2018

How to recursively zip a folder and its subfolders and add password protetion+encryption

Below I am compressing all files and subfolders in the folder /home/oracle/outputfiles:

cd /home/oracle/outputfiles
zip -r --encrypt myzipfile.zip *

You will be prompted for a password, which has to be verified.
If you are located in the parent directory, use
zip -r -q myzipfile mydir
where myzipfile is the name of the resulting zip file, and mydir is the name of the directory.
The .zip extension will be added to myzipfile automatically.

Tuesday, February 27, 2018

How to install APEX in an existing database


I have based my receipe on Tim Hall's instructions found at www.oracle-base.com, and I do not want to take credit for this. But the installation steps was executed by myself, and my notes may come in handy for others.

In my case, a developer needed access to the APEX_JSON package, and as confirmed with Oracle Development, the package is not available outside the APEX software stack. However, when APEX is installed in an existing database, you can start using the functionality in this package even if you don't need anything else that comes with APEX.

Here is how I performed the minimal installation:


1. Create a dedicated tablespace:

create bigfile tablespace apex datafile '/u02/oradata/testdb01/apex.dbf' size 128M autoextend on next 32M maxsize unlimited;

2. Download the latest APEX installation bundle from http://www.oracle.com/technetwork/developer-tools/apex/downloads/index.html#close, and upload it to your server

3. Unzip the file

4. change directory to the apex folder

5. Install the software.
The header of the apexins.sql file states its usage:
Arguments:
Position 1: Name of tablespace for Application Express application user
Position 2: Name of tablespace for Application Express files user
Position 3: Name of temporary tablespace or tablespace group
Position 4: Virtual directory for APEX images

sqlplus / as sysdba @apexins.sql APEX APEX TEMP /i/

6. Set the admin password:

sqlplus / as sysdba @apxchpwd.sql

7. Create the APEX_LISTENER and schema APEX_REST_PUBLIC_USER:

sqlplus / as sysdba @apex_rest_config.sql

There are additional, optional steps to be configured, depending on your needs. See this article at Oracle-base.com