Active Data Guard and Undo and ora-01555

Similar Questions

• Problem link DB between active Data Guard and reports application database

  My version of the 11.2.0.2.0 and OS database is Oracle Solaris 10 9/10.
  I am facing a problem in my custody of data Active data base for purposes of tax. Active Data guard information is as below.
  
  SQL > select name, database_role, open_mode from v$ database;
  
  NAME DATABASE_ROLE OPEN_MODE
  --------- ---------------- --------------------
  ORCL PHYSICS READ SHALL ONLY APPLY
  
  Detail of the problem is less than
  ------------------------------
  I have created a db link (name: DATADB_LINK) between active data guard and report of application of data base for purposes of tax.
  SQL > create database DATADB_LINK link to connect to HR identified by HR using 'DRFUNPD ';
  Database link created.
  
  But when I run a query using db link to my database of enforcement report I got this error below.
  
  ORA-01555: snapshot too old: rollback segment number 10 with the name ' _SYSSMU10_4261549777$ ' too small
  ORA-02063: preceding the line of DATADB_LINK
  
  Then I see logfile named database alart Active Data Guard and get below error
  
  ORA-01555 caused by the following SQL statement (SQL ID: 11yj3pucjguc8, time of request = 1 sec, SNA: 0x0000.07c708c3): SELECT "A2". "' BUSINESS_TRANSACTION_REFERENCE ', 'A2 '. "' BUSINESS_TRANSACTION_CODE ', MAX (CASE 'A1'. "TRANS_DATA_KEY"WHEN "feature' AND 'A1'." " END OF TRANS_DATA_VALUE"), MAX (CASE 'A1'. "TRANS_DATA_KEY" WHEN 'otherFeature' THEN 'A1' '. "" END OF TRANS_DATA_VALUE")
  
  But the interesting point if I run the query report directly in the Active Data Guard database, I got never error.
  
  So it's a problem of link DB between active Data Guard and other databases?

  Fazlul Kabir Mahfuz wrote:
  My version of the 11.2.0.2.0 and OS database is Oracle Solaris 10 9/10.
  I am facing a problem in my custody of data Active data base for purposes of tax. Active Data guard information is as below.

  SQL > select name, database_role, open_mode from v$ database;

  NAME DATABASE_ROLE OPEN_MODE
  --------- ---------------- --------------------
  ORCL PHYSICS READ SHALL ONLY APPLY

  Detail of the problem is less than
  ------------------------------
  I have created a db link (name: DATADB_LINK) between active data guard and report of application of data base for purposes of tax.
  SQL > create database DATADB_LINK link to connect to HR identified by HR using 'DRFUNPD ';
  Database link created.

  But when I run a query using db link to my database of enforcement report I got this error below.

  ORA-01555: snapshot too old: rollback segment number 10 with the name ' _SYSSMU10_4261549777$ ' too small
  ORA-02063: preceding the line of DATADB_LINK

  Then I see logfile named database alart Active Data Guard and get below error

  ORA-01555 caused by the following SQL statement (SQL ID: 11yj3pucjguc8, time of request = 1 sec, SNA: 0x0000.07c708c3): SELECT "A2". "' BUSINESS_TRANSACTION_REFERENCE ', 'A2 '. "' BUSINESS_TRANSACTION_CODE ', MAX (CASE 'A1'. "TRANS_DATA_KEY"WHEN "feature' AND 'A1'." " END OF TRANS_DATA_VALUE"), MAX (CASE 'A1'. "TRANS_DATA_KEY" WHEN 'otherFeature' THEN 'A1' '. "" END OF TRANS_DATA_VALUE")

  But the interesting point if I run the query report directly in the Active Data Guard database, I got never error.

  So it's a problem of link DB between active Data Guard and other databases?

  Check this statement that applies to your environment

  * ORA-01555 on Active Data Guard Standby Database [1273808.1 ID] *.

  also

  http://asktom.Oracle.com/pls/asktom/f?p=100:11:0:P11_QUESTION_ID:8908307196113

• Active Data Guard

  Hello
  Can we create an Active Data Guard in Oracle 10 g R2 release?
  
  
  Vinita

  Hi Vinita;

  Can we create an Active Data Guard in Oracle 10 g R2 release?

  You can not, you need to be level 11g. Please see:
  http://www.databasejournal.com/features/Oracle/article.php/3834931/using-Oracle-11gs-Active-Data-Guard-and-snapshot-standby-features.htm
  http://en.Wikipedia.org/wiki/Oracle_Data_Guard

  Respect of
  HELIOS

• Digital switchover and Active Data Guard

  I spin a quarter-Exadata V2 support systems (two instances CARS) and am changing my sleep Physics (an Exadata V1 rack with four RAC instances) in an Active Data Guard configuration (i.e. bearer Real - Time Query). The RDBMS runs 11.2.0.2 with BP7.
  
  My question is if I need to disable RTQ/Active Data Guard before a planned failover? If this is not the case, the transition process will automatically start QTR on the 'new' standby (i.e. my 'old' primary)? We'll test this on next month, but he was curious to know what to expect in advance.
  
  I use DGMGRL (Data Guard Broker) to manage the transition to digital.
  
  Any thoughts?
  
  Mark

  If you use the broker, as you must do with RAC, the broker must manage everything seamlessly. If it's not, and you need to test this, open an SR with Oracle's Support.

• Enable and disable Active Data Guard

  We are to build a simple primary (no CARS) to the physical environment pending (no CARS) Data Guard.
  Build us local and then shipment to the customer.
  During our generation we load data and the physical database in Active Data Guard 'mode '.
  To make it easier to check that the data is "protected".
  
  How to get back to "regular" managed recovery Active Data Guard "mode"?
  We were closing the physical Standby and then reopen in managed recovery "ordinary."
  Is it possible to do WITHOUT closing the physical Standby?
  
  Published by: user10159866 on April 22, 2010 12:15

  N ° the only way is at the stop of the physics of the day before and then do a START-UP SUPPORT and restart the application. What version do you use?

  Larry

  PS I guess you have Active Data Guard licensed? :^)

• Active Data Guard on DB 11 g and ADF business components

  Does anyone has experiences using business against a data source Active Data Guard components? Y at - it something ADF BC specific, you might need to know/set up regarding the parameters of the Application Module. Of course, we would read only your with her. Sorry it kinda turns out to be a beautiful finished question just want to know if it's a go or one thing more than any other possiblity.

  TheOtherDave,

  It should work perfectly. One thing you will need to know is the passivation of State mechanism used by ADF BC can use a store based on files or DB-table-based store. Obviously, if you want to use the DB-table-based store, you will have to configure it to point to a database where you can write to DB tables, which will be probably different from your database pending of DG. Otherwise, I can't think of specific issues (but I don't him have not personally tested).

  John

• Data Guard and Flashback

  I will be an active data guard with a primary instance and a secondary instance. I'm in the environment no CARS. My primary databases and Eve will be the flashback database on. I duration of blowback for one day, i.e., I should be able to flashback to the top of any time in the last 24 hours.

  When I do the RMAN backup on primary and remove the archiving log by using the rman command, RMA will not remove newspapers to archive that will be required for the flashback database (24 hours a day)?

  When I am applying newspapers archived on the backup site, I want to remove the archived logs that have been applied, but not required for the flashback database. How can I determine that? In the worst case scenario, I can always keep archived logs from 1 day to wait?

  1. Thank you.

  You need not remove the flashback logs.  Oracle automatically manages the removal of flashback logs.

  Configure retention policy to remove the archive logs more at some point.

  Configure archivelog deletion policy to manage the removal of logs that have been applied to the wait.

  for example:

  CONFIGURE RETENTION POLICY TO RECOVERY OF 7-DAY WINDOW;

  CONFIGURE ARCHIVELOG DELETION POLICY TO APPLIED ON STANDBY;

  In the worst case if you discover you need archivelogs that have been removed from the drive, then you can restore them because they have already been backed up by RMAN.

• Removal of RMAN, Data Guard, and Archives journal

  Our DG Environment running Oracle 11 g R2
  We have a 3 node DG Environment with
  A primary being
  B and C being Active Data Guard Standby
  Backups are taken out of B and go directly to tape.
  Standby Redo Logs and fast recovery area are used
  
  
  
  Holding the recommendation of "Using Recovery Manager with Oracle Data Guard in Oracle Database 10g"
  RMAN primary control ('A')
  CONFIGURE ARCHIVELOG DELETION POLICY TO APPLIED PENDING
  
  RMAN standby ("B") where the backup is performed
  CONFIGURE THE NONE ARCHIVELOG DELETION POLICY
  
  RMAN set the other Standby ("C")
  CONFIGURE ARCHIVELOG DELETION POLICY TO APPLIED PENDING
  
  How can we know what newspaper Archives are eligible to be removed from 'A' and 'C '?
  When remove it takes place?
  How can say us when archiving logs are deleted from the 'A' and 'C '?

  Dear user10260925,

  The documentation that you read is reliable but not enough.

  The Oracle can handle the archivelog directory and know that you are eligible for deletion. Stuff you posted here was derived from the books online and is supported and can be used when the Oracle knows and manages the archivelogs. This is simply called the flash recovery area. Please read the FRA at this very moment.

  Under the personal component normal circumstances in the industry use a few scripts to achieve on backup archivelog deletion.

  Here's a useful example to you;

  # Remove old archivelogs
  ######################################
  00,30 * * * * /home/oracle/scripts/delete_applied_redo_logs_OPTSTBY.sh
  ######################################
  
  vals3:/home/oracle#cat /home/oracle/scripts/delete_applied_redo_logs_OPTSTBY.sh
  export ORACLE_SID=optstby
  export ORACLE_HOME=/oracle/product/10.2.0/db_1
  cd /db/optima/archive/OPTPROD/archivelog
  
  /oracle/product/10.2.0/db_1/bin/sqlplus "/ as sysdba" @delete_applied_redo_logs.sql
  grep arc delete_applied_redo_logs.lst > delete_applied_redo_logs_1.sh
  chmod 755 delete_applied_redo_logs_1.sh
  sh delete_applied_redo_logs_1.sh
  rm delete_applied_redo_logs_1.sh
  rm delete_applied_redo_logs.lst
  

  vals3:/home/oracle#cd /db/optima/archive/OPTPROD/archivelog
  vals3:/db/optima/archive/OPTPROD/archivelog#cat delete_applied_redo_logs.sql
  set echo off
  set heading off
  spool /db/optima/archive/OPTPROD/archivelog/delete_applied_redo_logs.lst
  select 'rm -f ' || name from v$archived_log where applied = 'YES';
  spool off
  exit
  vals3:/db/optima/archive/OPTPROD/archivelog#
  

  Hope that helps.

  Ogan

• Data Guard and connectable to database

  Possible to have some project documents in a CDP that is used for Data Guard and other PDB files don't not ITIS?

  No, not in the current version. You must use the PDBs als for DG or nothing.

• Data Guard and auditor of the Apex

  Hi all

  I seem to be unable to understand the configuration of parameters for protective Apex listener when connecting to a database.

  We have a lot of servers of databases (without data guard) and we use the command line Wizard to generate the configuration file. It's straigthforeward and
  easy. We enter the server, port, and Service_name and it works.

  But for a guard database, there are two servers and I seem to be unable to set it in the right way. After doing some research I came across the parameter apex.db.customURL which should solve the problem.

  I removed the references to the server, port, and Service_name and put the key in.

  The result were errors of connection due to some incorrect port settings.

  SEVERE: The pool named: apex is not properly configured, error: IO error: format invalid number for the port number

  oracle.dbtools.common.jdbc.ConnectionPoolException: the pool named: apex is not properly configured, error: IO error: format invalid number for the port number

  at oracle.dbtools.common.jdbc.ConnectionPoolException.badConfiguration(ConnectionPoolException.java:65)

  What Miss me?

  Thank you
  Michael

  (Here's the rest of our configuration :)

  <? XML version = "1.0" encoding = "UTF-8" standalone = 'no '? >

  < ! DOCTYPE SYSTEM property "http://java.sun.com/dtd/properties.dtd" > ""

  Properties of <>

  < comment > saved on Mon Oct 19 18:28:41 CEST 2015 < / comment >

  < key "debug.printDebugToScreen entry" = > false < / entry >

  < key "security.disableDefaultExclusionList entry" = > false < / entry >

  < key = "db.password entry" > @055EA3CC68C35F70CF34A203A8EE1A55D411997069F6AE9053B3D1F0B951D84E0E < / entry >

  < key = "enter cache.maxEntries" > 500 < / entry >

  < key = "enter error.maxEntries" > 50 < / entry >

  < key = "enter security.maxEntries" > 2000 < / entry >

  < key = "cache.directory entry" > / tmp/apex/cache < / entry >

  < enter key = "jdbc. DriverType"> thin < / entry >

  < key = "enter log.maxEntries" > 50 < / entry >

  < enter key = "jdbc. MaxConnectionReuseCount"> 1000 < / entry >

  < key "log.logging entry" = > false < / entry >

  < enter key = "jdbc. InitialLimit' > 3 < / entry >

  < enter key = "jdbc. MaxLimit' 10 > < / entry >

  < key = "enter cache.monitorInterval" 60 > < / entry >

  < key = "enter cache.expiration" > 7 < / entry >

  < key = "enter jdbc.statementTimeout" > 900 < / entry >

  < enter key = "jdbc. MaxStatementsLimit' 10 > < / entry >

  < key = "misc.defaultPage entry" > apex < / entry >

  < key = "misc.compress" / entry >

  < enter key = "jdbc. MinLimit"> 1 < / entry >

  < key = "cache.type entry" > lru < / entry >

  < key "cache.caching entry" = > false < / entry >

  < key "error.keepErrorMessages entry" = > true < / entry >

  < key = "cache.procedureNameList" / entry >

  < key = "cache.duration entry" > days < / entry >

  < enter key = "jdbc. InactivityTimeout"1800 > < / entry >

  < key "debug.debugger entry" = > false < / entry >

  < key = "enter db.customURL" > JDBC: thin: @(DESCRIPTION = (ADDRESS_LIST = (ADRESSE = (COMMUNAUTÉ = tcp.world) (PROTOCOL = TCP) (host = DB-ENDUR) (Port = 1520)) (ADDRESS = (COMMUNITY = tcp.world)(PROTOCOL=TCP) (Host = DB-ENDURK) (PORT = 1521)) (LOAD_BALANCE = off)(FAILOVER=on)) (CONNECT_DATA = (SERVICE_NAME = ENDUR_PROD.) VERBUND.CO. «"" AT))) < / entry >»»»

  < / properties >

  
  
  

  Hi Michael Weinberger,

  Michael Weinberger wrote:

  I seem to be unable to understand the configuration of parameters for protective Apex listener when connecting to a database.

  We have a lot of servers of databases (without data guard) and we use the command line Wizard to generate the configuration file. It's straigthforeward and
  easy. We enter the server, port, and Service_name and it works.

  But for a guard database, there are two servers and I seem to be unable to set it in the right way. After doing some research I came across the parameter apex.db.customURL which should solve the problem.

  I removed the references to the server, port, and Service_name and put the key in.

  Keep the references the server name, port and service. No need to delete.

  The result were errors of connection due to some incorrect port settings.

  SEVERE: The pool named: apex is not properly configured, error: IO error: format invalid number for the port number

  oracle.dbtools.common.jdbc.ConnectionPoolException: the pool named: apex is not properly configured, error: IO error: format invalid number for the port number

  at oracle.dbtools.common.jdbc.ConnectionPoolException.badConfiguration(ConnectionPoolException.java:65)

  What Miss me?

  You must create two entries in the configuration file "defaults.xml" for your ADR (formerly APEX Listener).

  One for db.connectionType and, secondly, for db.customURL, for example:

  customurl
  jdbc:oracle:thin:@(DESCRIPTION=(ADDRESS_LIST=
  (ADDRESS=(COMMUNITY=tcp.world)(PROTOCOL=TCP)(HOST=DB-ENDUR)(Port = 1520))
  (ADDRESS=(COMMUNITY=tcp.world)(PROTOCOL=TCP)(HOST=DB-ENDURK)(PORT = 1521))
  (LOAD_BALANCE=off)(FAILOVER=on))(CONNECT_DATA=(SERVICE_NAME=ENDUR_PROD.VERBUND.CO.AT)))
  

  Reference: http://docs.oracle.com/cd/E56351_01/doc.30/e56293/config_file.htm#AELIG7204

  NOTE: After you change the configuration file, don't forget to restart independent ADR / Java EE application server support if ADR is deployed on one.

  Also check if your URL for a JDBC connection is working properly and if there are any questions, you can turn on debugging for ADR:

  Reference:

  • http://docs.Oracle.com/CD/E56351_01/doc.30/e56293/trouble.htm#AELIG7206
  • http://docs.Oracle.com/CD/E56351_01/doc.30/e56293/trouble.htm#AELIG7207

  Directed by Tony, you should post the ADR related questions to the appropriate forum. Reference: ADR, SODA & JSON in the database

  You can also move this thread on the forum of the ADR.

  Kind regards

  Kiran

• Active data guard lost synchronization

  Hi guys,.

  I hope you can give me some advice.

  I have a BD PRIMARY in a server and a STANDBY in another server.

  I do not configure these databases and I don't have too much experience with data protection.

  The problem is that the STAND BY lost the sincronization. I just watched a few views on the database and I found that the media database is READ only, so I think that is an Active Data Guard.

  -PRIMARY:

  SQL > SELECT name, State, database_role, open_mode connections, log_mode FROM v$ instance, v$ database;

  STATUS DATABASE_ROLE LOG_MODE CONNECTIONS

  --------- ---------------- -------------------- ------------

  BD PRIMARY READ WRITE ARCHIVELOG

  -STANDBY TIME:

  SQL > SELECT name, State, database_role, open_mode connections, log_mode FROM v$ instance, v$ database;

  STATUS DATABASE_ROLE LOG_MODE CONNECTIONS

  --------- ---------------- -------------------- ------------

  PHYSICAL STANDBY BD READ ONLY WITH APPLY ARCHIVELOG

  
  

  -PRIMARY:

  "SQL > select SEQUENCE #, FIRST_TIME NEXT_TIME, APPLIED, ARCHIVES of V$ ARCHIVED_LOG, whose name = ' STANDBY' FIRST_TIME agenda;

  
  

  SEQUENCE # FIRST_TI ARC APPLIED NEXT_TIM

  ---------- -------- -------- --------- ---

  56 18/03/15-18/03/15 NO YES

  57 18/03/15-18/03/15 NO YES

  58 18/03/15-18/03/15 NO YES

  59 18/03/15 03/19/15 NO YES

  60 19/03/15 03/19/15 NO YES

  61 19/03/15 03/20/15 NO YES

  ...

  426 13/05/15 13/05/15 NO YES

  427 13/05/15 13/05/15 NO YES

  428 13/05/15 13/05/15 NO YES

  429 13/05/15 05/14/15 NO YES

  430 14/05/15 05/15/15 NO YES

  I saw the newspaper archive all copy them correctly on the standby server, but they apply. I don't know why.

  How can I get the two database will be synchronized? What option I have?

  Best regards...

  Hello, you should check if Fal_server and Fal_client settings are configured in the main bodies and relief.

  In the standby instance fal_server must point to name PRIMARY, as it appears in tnsnames.ora

  fal_client must point in standby mode, as it appears in tnsnames.ora

  In the first instance, fal_server must point to the name of Eve, as it appears in tnsnames.ora, fal_client pointing to primary name, as it appears in tnsnames.ora

  on the other hand, you must restore the automatic sequence 33-55 on the backend and fal then transport and then primary eve because of the resolution of the gap.

  http://docs.Oracle.com/CD/B19306_01/server.102/b14239/log_transport.htm#i1268281

• Refresh the MV of standby Active Data Guard

  Is it possible to refresh a materialized view (quick or full) come from a standby database Active Data Guard open in read mode only w/apply?  Or can I just do the update of the primary?

  Hello mseberg,

  You made a very good point about the definition of a database of the READER.

  After you have answered, I found in MOS: create a distributed materialized view based on a master table on a database in read-only (Doc ID 276023.1)

  This explains that it is NOT possible to make a fast refresh of a database shall RO.  However, it *peut* be possible to do a complete refresh of a standby Number.  It depends on the query used to define the MV.  If the RDBMS Analyzer determines the MV is *able* to fast refresh (even if there may not need it right now), so you can't do a full refresh since the eve of the RO.  If the parser determines that the request is NOT able to fast refresh, then you can do a full refresh since the eve of the RO.  I tested this script on 11.2.0.3.8 and found to be true:

  CREATE THE MIMI MATERIALIZED VIEW. MV_TEST

  TABLESPACE USERS

  IMMEDIATE CONSTRUCTION

  COMPLETE REFRESH

  AS (SELECT * FROM TEST@MIMI_ROSTBY);

  ORA-16000: database opened for read-only access

  ORA-06512: at "SYS." DBMS_SNAPSHOT_UTL', line 364

  ORA-06512: at "SYS." DBMS_SNAPSHOT_UTL', line 446

  ORA-06512: at line 1

  Now, I've changed the MV query to use a 'one time' feature (ROWNUM), which causes the query become complex (not able to fast refresh).  I can now create a MV on the day before RO, because the query is no longer able to fast refresh and requires no registration on the eve of the RO:

  CREATE THE MIMI MATERIALIZED VIEW. MV_TEST

  TABLESPACE USERS

  IMMEDIATE CONSTRUCTION

  COMPLETE REFRESH

  AS (SELECT * FROM TEST@MIMI_ROSTBY WHERE ROWNUM > 0);

  Created.

• Oracle11g R2 Active Data guard using ASM problem?

  I configured oracle11g r2 RAC on 2 notes with the ASM (OS, unix) grid.
  ACR is running.
  Now I'm setting up Active data Guard.
  
  Under Network instance user + ASM and the listener is running.
  User is running under oracle listener static.
  Any drive is mounted.
  
  Oracle RAC and Data Guard directory and structure I've kept even.
  
  
  Now my problem is below:
  
  $. / rman target sys/HPinvent123nbl@dcpdb sys/HPinvent123nbl@drpdb ASSISTANT
  
  Recovery Manager: release 11.2.0.1.0 - Production on Wed Jan 16 16:28:32 2013
  
  Copyright (c) 1982, 2009, Oracle and/or its affiliates. All rights reserved.
  
  connected to target database: DCPDB (DBID = 316773134)
  connected to the auxiliary database: DRPDB (unassembled)
  
  RMAN > database of target duplicate for the eve of the active database.
  
  From 16 January 13 Db double
  using the control file of the target instead of recovery catalog database
  allocated channel: ORA_AUX_DISK_1
  channel ORA_AUX_DISK_1: SID = 5644 device type = DISK
  
  content of Script memory:
  {
  backup that copy reuse
  TargetFile ' / u02/app/oracle/product/11.2.0/dbhome_1/dbs/orapwdcpdb1' auxiliary format
  ' / u02/app/oracle/product/11.2.0/dbhome_1/dbs/orapwdrpdb';
  }
  execution of Script memory
  
  From backup 16 January 13
  allocated channel: ORA_DISK_1
  channel ORA_DISK_1: SID = 1897 instance = dcpdb1 = DISK device type
  Backup finished on 16 January 13
  RMAN-00571: ===========================================================
  RMAN-00569: = ERROR MESSAGE STACK FOLLOWS =.
  RMAN-00571: ===========================================================
  RMAN-03002: failure of Db in dual at 16/01/2013 16:28:48
  RMAN-06136: the auxiliary database ORACLE error: ORA-00200: control file could not be created
  ORA-00202: control file: '+ data '.
  ORA-17502: cannot create the file + data ksfdcre:4
  ORA-15001: 'DATA' diskgroup do not exist or are not mounted
  ORA-15055: unable to connect to the ASM instance
  ORA-01031: insufficient privileges
  
  RMAN >
  
  Please help. \
  
  
  Thank you
  Solaiman

  Is your Oracle asmdba group user name?

  What is the output of command mentioned below

  ID Oracle

• Active data guard for EBS

  Hi all

  EBS r12.2.4

  rhel6.5

  Active Dataguard applies in ebs?

  It is available in 11.1 or 11.2 only and up to?

  Thank you very much.

  JC

  Hello JC,.

  See the following blog: https://blogs.oracle.com/stevenChan/entry/oracle_active_data_guard_111

  See also ""E-Business Suite Oracle Active Data Guard Certifications ". "

  Coco - M

• Insert into a primary database of Eve Active Data Guard

  I use Oracle EE 11.2.0.3.8 on Solaris 10.

  I read that it is possible to do an insert in my database for the standby Active Data Guard by redirecting the insert to get on the primary via DB link.  I GLOBAL_NAMES set to TRUE and he needs to stay that way.  This means that I can't create a DB link pointing to my first as such:

  CREATE DATABASE LINK PROD. WORLD WITH THE HELP OF "PROD_PRIM";

  How can I get around this in order to insert data in my standby ADG primary?

  Hello;

  There is an article about it here by Uwe Hesse:

  http://uhesse.files.WordPress.com/2012/03/realtimequery.PDF

  I understand that if GLOBAL_NAMES is true, then, the database link must have the same name as the database in which it connects.

  Best regards

  mseberg