Good Read for PostgreSQL DBA

PostgreSQL 9 High Availability Cookbook

Link to Amazon

Good PostgreSQL HA guide

Several good guidelines for designing PostgreSQL databases and HA Clusters.
For example recipes for counting storage size, IOPS, cpu, memory etc...
Also very specific guides how to use different replications and their monitoring + management and HA solutions.

ORAchk to check databases (single and RAC)

Oracle ORAchk utility will replace RACchk utility and allow you to check also single instance databases. You can also send notification emails throught it. And schedule automatic check via cron jobs.

Here is quick guide to install and use ORAchk utility:

1. Download from My Oracle Support (you'll need MOS account for this):
Doc ID 1268927.2

NOTE! here you can also find user guide pdf!

2.  Install as oracle user in host you want to check:

      - unzip orachk.zip

3. Set notifications Email and automatic cron checks:
      - ./orachk -set "AUTORUN_SCHEDULE=3 1 * *; NOTIFICATION_EMAIL=<your_email_address>"

NOTE! Outlook (+ virus scanner) may not work proberly with these notification emails. So you might need some testing to get this work with it.
cron parameters in this set are (left to right):  hour (0 - 23), day of month (1 - 31), month (1 - 12), day of week (0 - 6) (0 to 6 are Sunday to Saturday)

4. Check settings:
./orachk -get all

5. Start daemon:
./orachk -d start

NOTE! with RAC this ask's root pwd for audit checks

6.  Stop daemon:
./orachk -d stop


7. To see all parameters:
./orachk -help

NOTE! you can also set auto restart. So script will be started automatically on node restarts.

Oracle DBA_TABLESPACE_USAGE_METRICS vs DBA_DATA_FILE with autoextent tablespaces

If you compare tablespace sizes with following two queries:

SQL> select * from dba_tablespace_usage_metrics

and

SQL> select x.tablespace_name,
    sum(x.bytes/1024/1024) "Total Space (Mb)" ,
    NVL((sum(x.bytes/1024/1024)-round(y.free,2)), sum(x.bytes/1024/1024)) "Used (Mb)",
     NVL(round(y.free,2), 0.00) "Free (Mb)",
     NVL(round(y.free/sum(x.bytes/1024/1024)*100,2), 0) "Free %"
from dba_data_files x,
    (select tablespace_name,sum(bytes/1024/1024) free
         from dba_free_space
         group by tablespace_name) y
where x.tablespace_name = y.tablespace_name (+)
group by x.tablespace_name,y.free;


They are looking very different especially if the tablespaces are using autoextent.
This is because :
Tablespace_size in DBA_TABLESPACE_USAGE_METRICS takes the maximum file size for autoextensible tablespace which corresponds to maxblocks in dba_data_files.
So basically DBA_TABLESPACE_USAGE_METRICS shows the maximum size to which a datafile can grow. And DBA_DATA_FILES (+ DBA_FREE_SPACE) current size of tablespace.
 

Look more from MOS (My Oracle Support) document:
Difference in Tablespace Size Values From dba_data_files and dba_tablespace_usage_metrics/V$Filespace_usage (Doc ID 455715.1)

Oracle Grid Control Agent autostart in RHEL / CentOS

When you install Grid Control Agent in server then the installer also create /etc/init.d/gcstartup script in the server. This init script wont automatically start Agent in server boot. But you can change it to do so. Do following as a root user.

1. Add following in the start of the script:

#!/bin/sh
# chkconfig: 35 10 90
# description: oracle grid control agent stop/start script on system up/down
#
# Use chkconfig command to start this script in appropriate run level.
# --------------------------------------------------------------------

#Source Function Library



2. And run following commands:

chkconfig --add gcstartup
chkconfig --levels 2345 gcstartup on
chkconfig --list gcstartup


This will work at least with Grid Control 11g agents and RHEL / Centos.

Oracle next extent check for tablespace fails and recyclebin (Grid Control)

If you using maintenance jobs to check does next extent fit in the tablespace and you get errors that they does not.

That means of course that your tablespace (or datafile) is so full that next extent would not fit into it.
"Quite like error ora-01653 unable to extend table in tablespace."

Sometimes this kind of errors can be confusing if you look tablespace usage from DBA_FREE_SPACE (or DBA_EXTENTS) . And this is because DBA_FREE_SPACE does not show dba_recyclebin usage. So if you remove database objects (for example table) from database but does not "purge" it then it will go into recyclebin and still takes same amount of space from same tablespace. This recyclebin used space can be taken into use if space is needed for new data. But it cannot be seen in DBA_FREE_SPACE as used space because it is reclaimable. But you can see this used space if you check for example tablespace (or datafile) usage from DBA_SEGMENTS.

Same thing is with the Grid Control (at least version 11G). If you check tablespaces usage it shows that there is free space to use in tablespace (this does not contain recyclebin space usage) but if you look the datafile space usage it will show real data space usage + recyclebin space usage.

NOTE! You wont get ORA-01653 errors in this kind of situations because Oracle can start to use space used by recyclebin when needed. But is is good to remember that you might get different space usage info if you wont purge dropped objects. 

MySQL MyISAM table repair.

MySQL MyISAM tables might break down from time to time.
You might get following errors in database error log:
--
[ERROR] mysqld: Incorrect key file
 for table '<TABLE_INDEX_FILE_NAME_AND_PATH>.MYI'; try to repair it

[ERROR] mysqld: Table '<TABLE_NAME_AND_PATH>' is marked as crashed and should be repaired--

If you get only error number from application then you can use perror utility from command prompt to get more info about error number (error numbers related to MyISAM tables usually are:
126 127 132 134 135 136 141 144 145) :

mysql/bin/perror 144

MySQL error code 144: Table is crashed and last repair failed



Check and fix MyISAM tables:

With myisamchk utility you can try to check and fix these tables/index files. Before run myisamchk  commands you need to stop MySQL server and go into database directory (or specify directory in the command). I prefer myisamchk because it garentees that no one is not trying to use tables when doing the repair (MySQL server need to be stopped when running it).


To Check table errors:
 
To check all tables (to check only one table specify that table name):
myisamchk *.MYI

To do slower and more extend-check check run:
myisamchk -e *.MYI

You can also check tables from SQL (when server is running):
CHECK TABLE <TABLE_NAME> <OPTIONS>;
CHECK TABLE <TABLE_NAME> QUICK;


To Fixing Tables or Index files:

Repairing tables might sometimes lead to lost of data so it is good practice to take backup of files or database before running these if it is possible.
 
Fix only those tables that are broken ( -q makes quick repair that repairs only index files):
mysql/bin/myisamchk -r -q <TABLE_NAME>


With big tables (and index files) and to get more performance for fixing use following options (you need to tune these values suitable for your environment) with myisamchk (to see all myisamchk options you can use myisamchk --help ):
mysql/bin/myisamchk --sort_buffer_size=2G \
           --key_buffer_size=2G \
           --read_buffer_size=512M \
           --write_buffer_size=512M \
      --tmpdir=/data/myisam_repair \
      --recover --quick \
     <TABLE_NAME>

You can also try to fix tables from SQL (when server is running) :
REPAIR TABLE <TABLE_NAME> <OPTIONS> ;
REPAIR TABLE <TABLE_NAME> QUICK;


NOTE: If some reason myisamchk did not fix your table/index files or you want to get more detailed guide and info about MyISAM table repair check these links:
http://dev.mysql.com/doc/refman/5.7/en/myisam-repair.html
http://dev.mysql.com/doc/refman/5.7/en/check-table.html
http://dev.mysql.com/doc/refman/5.1/en/repair-table.html

Oracle ORA-12564 TNS:connection refused ( shared server )

If you get "ORA-12564: TNS:connection refused" errors from your database connections.
Reason is typically misspelling in tnsnames.ora file.

But if you are using shared server connections these errors can be seen also if you have not enough dispatchers or your dispatcher max session limit is reached.

Usually with ORA-12564 you can also find these kind of errors in your listener logs:
"TNS-12520: TNS:listener could not find available handler for requested type of server"

This way you can check your connections via listener (this shows both dedicated and dispatcher connections):
lsnrctl services

With following sql you can check your dispatcher settings (these settings can be changed with ALTER SYSTEM SET dispatchers= ... commands):
SQL> select * from V$DISPATCHER;
SQL> select * from V$DISPATCHER_CONFIG;


If you does not get these errors in normal usage but only in occasionally then you also might want to check which users are doing most connections when this problem is on. You can do it with this sql (this uses RAC gv$session view so it shows cluster all instances connections. If you want only one instance connections you can use v$session):
SQL>  select INST_ID, USERNAME, count(SID) from gv$session group by USERNAME, INST_ID order by count(SID);

Oracle "opidcl aborting process unknown ospid (xxx) as a result of ORA-28" error.

If you get following ORA errors in your database alert log:
"opidcl aborting process unknown ospid (xxxx) as a result of ORA-28"

This usually means that some privileged user (dba) has killed sessions from database.

But if you see these errors a lot or all the time then this can be bug. This bug is affected
older (Oracle 11.1.0.6 and 11.1.0.7) versions. With bug error messages look like this:
"ORA-28 : opiodr aborting process unknown ospid (xxxxx_xxxxxxxxxxx) "

NOTE! If some application or user kill several (for example all one user) sessions at the same time there can be several errors in a row in alert log and it is still a normal situation.

Oracle 12c timestamps for datapump output file and console.

Starting on Oracle 12c you can set timestamps on for datapump (expdp and impdp) output file and console messages. With new LOGTIME option you can control timestamps printing. This option can have 4 different values (NONE, STATUS, LOGFILE, ALL).

NONE is default value. With it there is no additional timestamps in the output file or in the console. 

STATUS With this value timestamps are printed in the console but not in the output file.

LOGFILE With this value timestamps are printed in output file but not in the console. 

ALL With this value timestamps are printed in the output file and in the console.

Oracle 12c nologging for impdp

There is very useful new feature in Oracle 12c impdp which you can use to get rid of logging when you are doing import.

For example big bulk imports you might want to take logging of (both table and index) this way archivelog disk is not getting full and you can save some time.  You can do this with impdp "TRANSFORM=DISABLE_ARCHIVE_LOGGING:Y" option.

If you want you can also choose to remove only table or index logging:
only table data logging off during import:
TRANSFORM=DISABLE_ARCHIVE_LOGGING:Y:TABLE
only index data logging off during import:
TRANSFORM=DISABLE_ARCHIVE_LOGGING:Y:INDEX 

Default value for this parameter is TRANSFORM=DISABLE_ARCHIVE_LOGGING:N which is doing normal logging during import.

Oracle ORA-00020: maximum number of processes (xxx) exceeded

If you get following 'ORA-00020' errors in database alert log:
--
ORA-00020: maximum number of processes (xxx) exceeded
 ORA-20 errors will not be written to the alert log for
 the next minute. Please look at trace files to see all
 the ORA-20 errors.
Process m000 submission failed with error = 20
--

Then you probably wont get into the database via sqlplus or any other tool. This is because you got same error when you are trying to connect into the database. Error means that you need to increase processes parameter value or find out what is using too many processes and fix the problem. Best way to handle this kind of situation is to stop application/applications that are using this database and then you can again connect into the database and increase needed parameter values (alter system set processes=500 scope=spfile;) and restart the database.

If you cannot stop the applications to free processes then you can try to stop database with following way:
Run following as oracle user into database server:
export ORACLE_SID=<database_name>
sqlplus -prelim / as sysdba

and then run 'shutdown immediate' or 'shutdown abort' . Then 'startup' or 'startup mount' and do the parameter change (after parameter change you need to do restart the database). But remember that this can damage your applications data (especially if you use abort which just stops the database right away.) So it is better to just stop the application/applications even it means little downtime.

NOTE! When you connect into database via 'sqlplus -prelim' then you can also try to check what is causing these errors with oradebug (hanganalyze). And if you find some problematic sessions/SQLs you can kill (via OS) just those without need to restart the whole database. But this can take some time so more quickly fix use the above instructions. More info about oradebug and hanganalyze can be find from My Oracle Support (MOS) documents:
215858.1
and
310830.1

Oracle 12c Invisible Columns

With Oracle 12c you can use Invisible Columns to hide table columns for your testing or for other purposes.
Column is invisible for following operations (but you can use normal DML operations for invisible columns):
1. SELECT * FROM <table_name>;
2. DESCRIBE  <table_name> (via sqlplus and OCI)
3. %ROWTYPE attribute declarations in PL/SQL

This way you use invisible columns:
For example:
CREATE TABLE test_table(
  test_id NUMBER,
  test_name VARCHAR2(32),
  starting_time TIMESTAMP,
  ending_time TIMESTAMP );


You can change columns to invisible and to visible:
ALTER TABLE test_table MODIFY (test_name INVISIBLE);
ALTER TABLE test_table MODIFY (test_name VISIBLE);

NOTE! When you set column invisible the column order changes (invisible column is removed from column order). And If you set same column back to visible it is placed last in table column order.


You can also add new columns with invisible on:
ALTER TABLE test_table ADD ( tester_id NUMBER INVISIBLE );


You can also create table with invisible columns. Just add INVISIBLE after column datatype:
CREATE TABLE test_table(
  test_id NUMBER INVISIBLE,
  test_name VARCHAR2(32),
  starting_time TIMESTAMP,
  ending_time TIMESTAMP );


You can make normal SQL DML operations for invisible column like this (for previous test table): INSERT INTO test_table (test_id, test_name, starting_time, ending_time) VALUES (1110, 'Test_1', '12-Oct-13', '15-Oct-13');



NOTE! The following types of tables cannot have invisible columns: External tables, Cluster tables, Temporary tables. Also attributes of user-defined types cannot be invisible.

You can find more info about Invisible Columns here:
Understand Invisible Columns

Oracle 12c Temporal Validity time periods in tables.

In Oracle 12c there is new feature called "Temporal Validity". With it you can create time periods between two columns and use these periods for queries.

Example:
-Create table with Temporal Validity period
(you can also add PERIOD FOR in existing table with "ALTER TABLE" clause):
CREATE TABLE test_table(
  test_id NUMBER,
  test_name VARCHAR2(32),
  starting_time TIMESTAMP,
  ending_time TIMESTAMP,
PERIOD FOR testing_time (starting_time, ending_time));

-Inserts are working just like before (PERIOD FOR is not column)
(There can be also NULL values if table constraints accept those.):
INSERT INTO test_table VALUES (1110, 'Test_1', '12-Oct-13', '15-Oct-13');
INSERT INTO test_table VALUES (1110, 'Test_1', '14-Oct-13', null);

- PERIOD FOR gives more variety for your queries (but you can also query table without it):
-This will return all rows that got given date in their time period (first example row):
SELECT * from test_table AS OF PERIOD FOR testing_time TO_TIMESTAMP('13-Oct-13');

-You can also use Period For in "BETWEEN" clause. This will return both example rows.:
SELECT * from test_table VERSIONS PERIOD FOR testing_time BETWEEN
TO_TIMESTAMP('13-Oct-13') AND TO_TIMESTAMP('16-Oct-13');


NOTE: Flashback Query has been extended to support queries on Temporal Validity dimensions. 

You can find more info about Temporal Validity from here:
Oracle 12c New Features
 and here:
Oracle 12c Desing Basics

Oracle ORA-7445 [kkzufst] errors in alertlog.

ORA-7445 [kkzufst] errors in alertlog are result from bugs in materialized views refresh jobs.

If you got these bugs you'll see these kind of errors in your database alertlog
(there is always ORA-07445 and kkzufst in error but other clauses can vary.):
--
Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x1000] [PC:0x44BBFCE, kkzufst()+111] [flags: 0x0, count: 1]
Errors in file...
ORA-07445 : exception encountered: core dump [kkzufst()+111] ...
Incident details in:...
--

In incident log you'll see something like this (here you can also see problematic SQL or PL/SQL clause in "Current SQL Statement ..." section)
(there is always ORA-07445 and kkzufst in error but other clauses can vary.):
--
Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x1000] [PC:0x44BBFCE, kkzufst()+111] [flags: 0x0, count: 1]
.
.
.
dbkedDefDump(): Starting a non-incident diagnostic dump (flags=0x3, level=3, mask=0x0)
----- Current SQL Statement for this session (sql_id=...

--


Fixes:
Bugs are more accurately 9554075 and 9656434
These both bugs are fixed in Oracle version 11.2.0.2 SPS (Server Patch Set).
And in Oracle version 12 fixes are in following versions: 9554075 -> version Oracle 12.1.0.1 (Base Release). 9656434 -> version Oracle 12.2 (Future release. (Not yet released))


More info about these bugs can be find from MOS (My Oracle Support) document:  
1288415.1

Oracle 12c enhancements for SCAN

There are new features for SCAN in Oracle 12c:

1. SCAN and Oracle Clusterware managed VIPs now support IPv6 based IP addresses.
So now we can use also IPv6 addresses with SCAN. If you check SCAN configs with:
srvctl config scan
it will list you more info: SCAN name, IPv4 address, IPv6 address, SCAN number+VIP



2. SCAN is by default restricted to only accept service registration from nodes in the cluster.
If you want to add nodes (for example for the database that are not using cluster's private interconnect) then you can do it with command:
srvctl modify scan_listener -invitednodes <new_database_name> –update
You can also run this modify before you present real node in the cluster.
 
If you check SCAN-listener configs with:
srvctl config scan_listener
it will list you more info: SCAN-listener name+port, Registration invited nodes, Registration invited subnets



3. SCAN supports multiple subnets in the cluster (one SCAN per subnet).

Before you can do this you must enable multiple subnets for the cluster (This is a post-installation task an cannot be made in Grid Infra installation). 

 


NOTE! Here you can find more info about SCAN and Oracle 12c enhancements :
http://www.oracle.com/technetwork/products/clustering/overview/scan-129069.pdf

Oracle 12c New Features for table Defaults

With Oracle 12c there is improvements in tables default values.

Starting on Oracle 12c you can use sequence nextval straight in column default or you can use "generated as identity" for column default (this will generate new sequence for table and start use it for column default values) This new sequence is linked with table so if you remove the table also the sequence is removed. You can also use "generated by default as identity" when you create new table sequence. This will allow you to override sequence value in inserts. If you use default word as column value or leave column and it's value out of the insert the sequence is used but you can also give new value for this column when "by default" is used.

With Oracle 12c you can also give default values for null columns. So if you are inserting null value into column that have for example this default settings: "default ON NULL 11" then 11 value will be inserted into table instead of null.

Examples:

Create table clause with using existing sequence:
SQL> CREATE TABLE TEST_TABLE ( id int seq_test_id.nextval primary key, name varchar2(32) );


Create table clause with generating new identity sequence:
SQL> CREATE TABLE TEST_TABLE ( id int generated as identity primary key, name varchar2(32) );

Create table clause with generating new identity sequence with sequence init values:
SQL> CREATE TABLE TEST_TABLE( id int generated as identity (start with 10000) primary key, name varchar2(32) );

Override sequence value when table squence is created with "generated by default as identity":
SQL> INSERT INTO TEST_TABLE (id, name) VALUES ( 10, 'test_name');

Using sequence value when table squence is created with "generated by default as identity":
SQL> INSERT INTO TEST_TABLE (id, name) VALUES ( default, 'test_name');
or
SQL> INSERT INTO TEST_TABLE (name) VALUES ( 'test_name');

Oracle 11.2.0.3 ORA-04030: out of process memory when trying to allocate xxxx bytes (kxs-heap-w,KGL Iterator information)

If you get the ORA-04030: ... (kxs-heap-w,KGL Iterator information) error in alert log
and if you see obsoleted parent cursors lying around.
Then you are hitting Bug 12791981 - ORA-4030 due to "KGL Iterator information" . 
Look more from MOS (My Oracle Support) document 12791981.8 .

If you are trying to check those obsolete cursors you are probably hitting the same error ORA-04030 and then you know that there are also those obsoleted cursors in the database.

With this SQL you can check those obsoleted cursors:
select sql_id, address, IS_OBSOLETE, count(*) from v$sql where IS_OBSOLETE='Y' group by sql_id, address, IS_OBSOLETE having count(*)>=10;

Or just:
select count(*) from v$sql where IS_OBSOLETE='Y'; 

You can fix this bug with patch 12791981 .


NOTE! This bug can also be fixed with patch 14799269 (this is more complete fix according to MOS) but it is only available for 11.2.0.3.1 version and newer ones.

Oracle RAC cluster NIC bonding.

You need to do following steps to start using NIC bonding with Oracle RAC (this add both public and private interconnect bond interfaces):

NOTE! This operation needs full downtime from cluster databases. And depending your environment you might need server reboots during this settings.

1. Add new bond0 and bond1 interfaces for RAC cluster globally via oifcfg. public interface is bond0 and cluster_interconnect interface bond1:

Get current network interface configuration being used by cluster as oracle user:
oifcfg getif




Set new bond interfaces – Updates OCR (-global make these changes to all nodes on cluster) run these in one node as oracle user (IPs can be same as before but interface name is changing):
oifcfg setif -global bond0/10.77.5.0:public
oifcfg setif -global bond1/10.37.24.0:cluster_interconnect
oifcfg getif

2. Stop databases and disable + stop crs on all nodes:

Do this on one node as oracle user (and all databases in cluster):
srvctl status database -d <database_name>  
srvctl stop database -d <database_name>
srvctl status database -d <database_name>  

Do these on both nodes as root user:
crsctl disable crs
crsctl check crs
crstcl stop crs
crsctl check crs
 

3.Change OS network interfaces to use NIC bonding like this (this guide only add bond0 but it is better to make bond0 (eth0 and eth1) for public and bond1 (eth2 and eth3 ) for interconnect. And you don't need to make alias interfaces.): 

http://www.oracle-base.com/articles/linux/nic-channel-bonding.php



If you need to reboot nodes because of your environment then do it now.

4. Enable and start crs on all nodes:

Do these on both nodes as root user:
crsctl enable crs
crsctl check crs
crstcl start crs
crsctl check crs
 check that crs started on both nodes run as root user on one node:
crsctl stat res -t

 

5. Remove old interfaces from RAC cluster via oifcfg:

Get current pvt. interconnect info as oracle user in one node:
oifcfg getif
 
Delete old interfaces eth0 and eth1 as oracle user in one node:
oifcfg delif -global eth0/10.77.5.0
oifcfg delif -global eth1/10.37.24.0

Check that only new bond interfaces are visible as oracle user in one node:
oifcfg getif

6. Repair scan- and vip- addresses using bond0 instead of eth0 :

Check cluster current network status as oracle user:
srvctl status nodeapps
Check current VIP settings. run this on all nodes as oracle user:
srvctl config vip -n <node_name>

Check current SCAN settings. run this on one node as oracle user:
srvctl config scan

Stop cluster nodeapps as oracle user:
srvctl stop nodeapps

Change SCAN/VIP settings. run this on all nodes as root user (set correct IP for all nodes):
srvctl modify nodeapps -n <node_name> -A 10.77.5.129/255.255.255.0/bond0

Check current VIP settings. run this on all nodes as oracle user:
srvctl config vip -n <node_name>

Start cluster nodeapps as oracle user:
srvctl stop nodeapps
Check cluster current network status as oracle user:
srvctl status nodeapps

Check current SCAN settings. run this on one node as oracle user:
srvctl config scan


 

7. Restart crs to see that it is starting correctly:

Do these on both nodes as root user: 

crsctl check crs
crstcl stop crs
crsctl check crs
crstcl start crs
crsctl check crs

8. Restart databases:

Do this on one node as oracle user (and all databases in cluster):
srvctl status database -d <database_name> 
srvctl start database -d <database_name> 
srvctl status database -d <database_name>

9. Test that you can connect into databases via all SCAN/VIP IPs. You can do this for example via sqlplus.

 

Oracle 11.2.0.3.0 DATABASE CRASHED DUE TO ORA-240 AND ORA-15064

There is a bug in Oracle 11.2.0.3.0 which can make your database instance restarting itself.
If you get following errors in database alert.log you know this bug is affecting your database:
ORA-00240: control file enqueue held for more than 120 seconds
ORA-29770: global enqueue process LCK0 (OSID 12329) is hung for more than 70 seconds
ORA-15064: communication failure with ASM instance

There is bugfix for this problem and you can download it from My Oracle Support (MOS) patch:
13914613

Other way to fix this is to update your database to newest version where this is also fixed

More info about this can be find from MOS documents:
Database Instance Crashes with ORA-15064 ORA-03135 ORA-00240 on 11.2 (Doc ID 1487108.1)

Bug 13914613 - Excessive time holding shared pool latch in kghfrunp with auto memory management (Doc ID 13914613.8) 


 

Oracle ORA-21561 : OID generation failed

If Oracle client connection is giving "ORA-21561 : OID generation failed" error.
Like this:


sqlplus <test_user>/<test_user_password>@<database_name>

.
.
.

ERROR:
ORA-21561 : OID generation failed

Enter user-name: <username>/<password> @ <tns connect string>


Then the problem is most likely in the client machine hosts file.
Check that there is client machine fully qualified name and short name in the client machine hosts file. If these are missing you'll get ORA-21561 errors when trying to connect server.