One of the database is down in streaming environment

I have scenario where one of the database in (two different database) streaming environment is unavailable due to nework/link/database maintenance purpose. Since my other database is up and running and DML/DDL operations are in process how can I avoid following error
SQL> insert into test values ('eee', 'code e');
1 row created.
SQL> commit;
commit
ERROR at line 1:
ORA-02050: transaction 5.23.8870 rolled back, some remote DBs may be in-doubt
ORA-02068: following severe error from OTHER_DOWN_DATABASE
ORA-03113: end-of-file on communication channel
============================================
Can I change some parameters in DBMS_STREAMS_ADM Package or any other packages to avoid above situation

Thanks for the reply.
I am getting the same error even If I just do select on the table which I have for streaming
SQL> select * from test@dbtest;
select * from test@dbtest
ERROR at line 1:
ORA-02068: following severe error from DBTEST ---destination database
ORA-03113: end-of-file on communication channel.
Q- Wich database is down? Is it the source database or the destination database?
For now I am testing with keeping down destination database.
Q- If it's the source db, just stop the capture. That's all you really have to do.
A- This can be good for planned maintenance, but how about unplanned outage or sudden problem (with networks/links or all of a sudden I can not stream my data changes to destination database).
Q--The error you're describing is unclear.... Do you get that error immediately when you issue the insert and the commit? That sound very very strange to me because streams does asynchronous replication... You would not get such an error for db down from my opinion.
A-- First time, I got this error when I updated a row and then tried to commit while my destination database was down.
Q- It looks like the problem has nothing to do with streams but rather a distributed transaction problem (2pc)...
A- If it is transaction problem should I be getting similar error while selecting from destination table, see the above error

Similar Messages

  • Based on a SQLException how to know if the database is down?

    Hi
    How can I know based on a SQLException that the database is down?. I know I can check the errorCode but which error codes indicate that the database is DOWN or not available?
    Regards,
    Néstor Boscán

    If the SQLException is IO exception: Connection refused, one of the causes is that the database instance has not been started. If the SQLException is
    IO exception: The Network Adapter could not establish the connection., one of the causes is that the database TNSListener has not been started.

  • Services of all the databases goes down

    Hi ,
    2 days back, Services of all the databases goes down on one of the PROD server.
    When i tried to start the services using SRVCTL command .
    I was getting the error below.
    oracle@shbaorc07b bin]$ srvctl status database -d EDW
    PRKH-1010 : Unable to communicate with CRS services.
    [Communications Error(Native: prsr_initCLSS:[3])]
    PRKO-2005 : Application error: Failure in getting Cluster Database Configuration for: EDWX
    PRKC-1056 : Failed to get the hostname for node + PRKH-1010 + prsr_initCLSS
    after this error, i came to know that there is some problem with the services.
    When i try to start the crs services, was not able to start the services
    as a root user user,
    I run the below command.
    crsctl start crs
    to my surprise the services of crs was not coming up,
    Finally I have to start the server and all the cluster services came up.
    I want to find out why this was happen, i need to find the bottle neck so this should not happen again as this is a PROD server and lots of prod database are residing on this server.
    Please find crsd.log below.
    2010-04-22 15:38:22.361: [  CRSRES][1539545408]0ora.shbaorc07b.LISTENER_UC4XPRD2.lsnr failed on shbaorc07b relocating.
    2010-04-22 15:38:22.397: [  CRSRES][1539545408]0Cannot relocate ora.shbaorc07b.LISTENER_UC4XPRD2.lsnrStopping dependents
    2010-04-22 15:38:22.407: [  CRSRES][1539545408]0StopResource: setting CLI values
    2010-04-22 15:38:27.318: [  CRSAPP][1539545408]0CheckResource error for ora.shbaorc07b.LISTENER_CTLGPRD2.lsnr error code = 1
    2010-04-22 15:38:27.321: [  CRSRES][1539545408]0In stateChanged, ora.shbaorc07b.LISTENER_CTLGPRD2.lsnr target is ONLINE
    2010-04-22 15:38:27.322: [  CRSRES][1539545408]0ora.shbaorc07b.LISTENER_CTLGPRD2.lsnr on shbaorc07b went OFFLINE unexpectedly
    2010-04-22 15:38:27.322: [  CRSRES][1539545408]0StopResource: setting CLI values
    2010-04-22 15:38:27.336: [  CRSRES][1539545408]0Attempting to stop `ora.shbaorc07b.LISTENER_CTLGPRD2.lsnr` on member `shbaorc07b`
    one more strange thing happen while restarting the 2nd server, the first server also restarted.
    Please help me regarding this?
    Thanks in advance.

    Hi,
    Thanks.
    please find the OS version
    Linux shbaorc07a.corp.shire.com 2.6.18-53.1.14.el5 #1 SMP Tue Feb 19 07:18:46 EST 2008 x86_64 x86_64 x86_64 GNU/Linux
    database version:
    10.2.0.4.0
    Node 2:
    [    CSSD]2010-04-22 19:41:19.438 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock IGDPAXPRDALL type 2
    [    CSSD]2010-04-22 19:41:19.438 [1273047360] >TRACE: clssgmCleanupOrphanMembers: cleaning up remote mbr(1) grock(IGDPAXPRDALL) birth(2/0)
    [    CSSD]2010-04-22 19:41:19.438 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock IGGMCXPRDALL type 2
    [    CSSD]2010-04-22 19:41:19.438 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock IGEASYPRDALL type 2
    [    CSSD]2010-04-22 19:41:19.438 [1273047360] >TRACE: clssgmCleanupOrphanMembers: cleaning up remote mbr(1) grock(IGEASYPRDALL) birth(2/0)
    [    CSSD]2010-04-22 19:41:19.438 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock IGCTRMPRDALL type 2
    [    CSSD]2010-04-22 19:41:19.438 [1273047360] >TRACE: clssgmCleanupOrphanMembers: cleaning up remote mbr(1) grock(IGCTRMPRDALL) birth(2/0)
    [    CSSD]2010-04-22 19:41:19.439 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock IGEDWXPRDALL type 2
    [    CSSD]2010-04-22 19:41:19.439 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock IGNUSBPRDALL type 2
    [    CSSD]2010-04-22 19:41:19.439 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock IGETLXPRDALL type 2
    [    CSSD]2010-04-22 19:41:19.439 [1273047360] >TRACE: clssgmCleanupOrphanMembers: cleaning up remote mbr(1) grock(IGETLXPRDALL) birth(2/0)
    [    CSSD]2010-04-22 19:41:19.439 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock IGCTXXPRDALL type 2
    [    CSSD]2010-04-22 19:41:19.439 [1273047360] >TRACE: clssgmCleanupOrphanMembers: cleaning up remote mbr(1) grock(IGCTXXPRDALL) birth(2/0)
    [    CSSD]2010-04-22 19:41:19.439 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock IGVVC30PRDALL type 2
    [    CSSD]2010-04-22 19:41:19.439 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock IGVUM30PRDALL type 2
    [    CSSD]2010-04-22 19:41:19.439 [1273047360] >TRACE: clssgmCleanupGrocks: cleaning up grock #CSS_CLSSOMON type 2
    [    CSSD]2010-04-22 19:41:19.439 [1273047360] >TRACE: clssgmCleanupOrphanMembers: cleaning up remote mbr(1) grock(#CSS_CLSSOMON) birth(2/0)
    [    CSSD]2010-04-22 19:41:19.445 [1273047360] >TRACE: clssgmEstablishConnections: 3 nodes in cluster incarn 5
    [    CSSD]2010-04-22 19:41:19.449 [1231087936] >TRACE: clssgmPeerListener: connects done (3/3)
    [    CSSD]2010-04-22 19:41:19.449 [1273047360] >TRACE: clssgmEstablishMasterNode: MASTER for 5 is node(3) birth(1)
    [    CSSD]2010-04-22 19:41:19.458 [1231087936] >TRACE: clssgmHandleDBDone(): src/dest (3/65535) size(72) incarn 5
    [    CSSD]CLSS-3000: reconfiguration successful, incarnation 5 with 3 nodes
    [    CSSD]CLSS-3001: local node number 2, master node number 3
    [    CSSD]2010-04-22 19:41:19.458 [1273047360] >TRACE: clssgmReconfigThread: completed for reconfig(5), with status(1)
    [    CSSD]2010-04-22 19:49:20.816 [1189128512] >TRACE: clssnmConnComplete: probe from node 1, your version: 10.2.1.2
    , support PENDINA: 1
    [    CSSD]2010-04-22 19:49:20.816 [1189128512] >TRACE: clssnmConnComplete: MSGSRC 1, type 5, node 1, flags 0x0001, con 0x2aaab01abb00, probe (nil)
    [    CSSD]2010-04-22 19:49:20.816 [1189128512] >TRACE: clssnmConnComplete: node 1, shbaorc07a, con(0x2aaab01abb00), probcon((nil)), ninfcon((nil)), node unique 1271962158, prev unique 0, msg unique 1271962158 node state 0
    [    CSSD]2010-04-22 19:49:20.816 [1189128512] >TRACE: clssnmsendConnAck: node 1, node state 0
    [    CSSD]2010-04-22 19:49:20.816 [1189128512] >TRACE: clssnmConnComplete: connecting to node 1 (con 0x2aaab01abb00), ninfcon (0x2aaab01abb00), state (0)
    [    CSSD]2010-04-22 19:49:20.816 [1189128512] >TRACE: clssnmConnComplete: connected to node 1 (con 0x2aaab01abb00), ninfcon (0x2aaab01abb00), state (0), flag (1037)
    [    CSSD]2010-04-22 19:49:20.927 [1262557504] >TRACE: clssnmDoSyncUpdate: Initiating sync 6
    [    CSSD]2010-04-22 19:49:20.927 [1262557504] >TRACE: clssnmDoSyncUpdate: diskTimeout set to (57000)ms
    [    CSSD]2010-04-22 19:49:20.927 [1262557504] >TRACE: clssnmSetupAckWait: Ack message type (11)
    [    CSSD]2010-04-22 19:49:20.927 [1262557504] >TRACE: clssnmSetupAckWait: node(1) is ALIVE
    [    CSSD]2010-04-22 19:49:20.927 [1262557504] >TRACE: clssnmSetupAckWait: node(2) is ALIVE
    [    CSSD]2010-04-22 19:49:20.927 [1262557504] >TRACE: clssnmSetupAckWait: node(3) is ALIVE
    [    CSSD]2010-04-22 19:49:20.927 [1262557504] >TRACE: clssnmSetupAckWait: node(4) is ALIVE
    [    CSSD]2010-04-22 19:49:20.927 [1262557504] >TRACE: clssnmSendSync: syncSeqNo(6)
    [    CSSD]2010-04-22 19:49:20.927 [1189128512] >TRACE: clssnmHandleSync: diskTimeout set to (57000)ms
    [    CSSD]2010-04-22 19:49:20.927 [1262557504] >TRACE: clssnmWaitForAcks: Ack message type(11), ackCount(4)
    [    CSSD]2010-04-22 19:49:20.927 [1189128512] >TRACE: clssnmHandleSync: Acknowledging sync: src[2] srcName[shbaorc07b] seq[1] sync[6]
    6911,1 99%
    node 1:
    [    CSSD]2010-04-22 18:56:00.996 [1189128512] >USER: clssnmHandleUpdate: SYNC(4) from node(4) completed
    [    CSSD]2010-04-22 18:56:00.996 [1189128512] >USER: clssnmHandleUpdate: NODE 1 (shbaorc07a) IS ACTIVE MEMBER OF CLUSTER
    [    CSSD]2010-04-22 18:56:00.996 [1189128512] >USER: clssnmHandleUpdate: NODE 2 (shbaorc07b) IS ACTIVE MEMBER OF CLUSTER
    [    CSSD]2010-04-22 18:56:00.996 [1189128512] >USER: clssnmHandleUpdate: NODE 3 (shbaorc07c) IS ACTIVE MEMBER OF CLUSTER
    [    CSSD]2010-04-22 18:56:00.996 [1189128512] >USER: clssnmHandleUpdate: NODE 4 (shbaorc07d) IS ACTIVE MEMBER OF CLUSTER
    [    CSSD]2010-04-22 18:56:00.996 [1189128512] >TRACE: clssnmHandleUpdate: diskTimeout set to (200000)ms
    [    CSSD]2010-04-22 18:56:01.008 [1273047360] >TRACE: clssgmReconfigThread: started for reconfig (4)
    [    CSSD]2010-04-22 18:56:01.008 [1273047360] >USER: NMEVENT_RECONFIG [00][00][00][1e]
    [    CSSD]2010-04-22 18:56:01.020 [1273047360] >TRACE: clssgmEstablishConnections: 4 nodes in cluster incarn 4
    [    CSSD]2010-04-22 18:56:01.096 [1231087936] >TRACE: clssgmInitialRecv: (0x2aaab4001310) accepted a new connection from node 2 born at 4 active (4, 4), vers (10,3,1,2)
    [    CSSD]2010-04-22 18:56:01.096 [1231087936] >TRACE: clssgmInitialRecv: conns done (4/4)
    [    CSSD]2010-04-22 18:56:01.096 [1273047360] >TRACE: clssgmEstablishMasterNode: MASTER for 4 is node(3) birth(1)
    [    CSSD]2010-04-22 18:56:01.138 [1231087936] >TRACE: clssgmHandleDBDone(): src/dest (3/65535) size(72) incarn 4
    [    CSSD]CLSS-3000: reconfiguration successful, incarnation 4 with 4 nodes
    [    CSSD]CLSS-3001: local node number 1, master node number 3
    [    CSSD]2010-04-22 18:56:01.138 [1273047360] >TRACE: clssgmReconfigThread: completed for reconfig(4), with status(1)
    [    CSSD]2010-04-22 19:40:00.531 [1231087936] >TRACE: clscsendx: (0x2aaab00fa3b0) Connection not active
    [    CSSD]2010-04-22 19:40:00.531 [1231087936] >TRACE: clssgmSendClient: Send failed rc 6, con (0x2aaab00fa3b0), client (0x2aaab00fa6b0), proc ((nil))
    [    CSSD]2010-04-22 19:40:19.410 [1231087936] >TRACE: clscsendx: (0x2aaab02acd00) Connection not active
    [    CSSD]2010-04-22 19:40:19.410 [1231087936] >TRACE: clssgmSendClient: Send failed rc 6, con (0x2aaab02acd00), client (0x2aaab02ad7c0), proc ((nil))
    [    CSSD]2010-04-22 19:40:19.410 [1231087936] >TRACE: clscsendx: (0x2aaab02ad570) Connection not active
    [    CSSD]2010-04-22 19:40:19.410 [1231087936] >TRACE: clssgmSendClient: Send failed rc 6, con (0x2aaab02ad570), client (0x2aaab02af660), proc ((nil))
    [    CSSD]2010-04-22 19:40:19.438 [1231087936] >TRACE: clscsendx: (0x2aaab0295780) Connection not active
    [    CSSD]2010-04-22 19:40:19.438 [1231087936] >TRACE: clssgmSendClient: Send failed rc 6, con (0x2aaab0295780), client (0x2aaab02859f0), proc ((nil))
    [    CSSD]2010-04-22 19:40:19.476 [1231087936] >TRACE: clscsendx: (0x2aaab029d150) Connection not active
    [    CSSD]2010-04-22 19:40:19.476 [1231087936] >TRACE: clssgmSendClient: Send failed rc 6, con (0x2aaab029d150), client (0x2aaab029dbb0), proc ((nil))
    [    CSSD]2010-04-22 19:40:19.476 [1231087936] >TRACE: clscsendx: (0x2aaab029d890) Connection not active
    [    CSSD]2010-04-22 19:40:19.476 [1231087936] >TRACE: clssgmSendClient: Send failed rc 6, con (0x2aaab029d890), client (0x2aaab029e1b0), proc ((nil))
    [    CSSD]2010-04-22 19:40:19.707 [1231087936] >TRACE: clscsendx: (0x2aaab02cdb30) Connection not active
    [    CSSD]2010-04-22 19:40:19.707 [1231087936] >TRACE: clssgmSendClient: Send failed rc 6, con (0x2aaab02cdb30), client (0x2aaab02cdf60), proc ((nil))
    [    CSSD]2010-04-22 19:40:19.710 [1231087936] >TRACE: clscsendx: (0x2aaab02ca9e0) Connection not active
    please help me on this.
    Thanks
    Edited by: user646034 on Apr 26, 2010 1:13 AM

  • How to create  a procedure to send a mail if the Database is down?

    Hi,
    I have created the below procedure to send a mail if the count is less than 1300. It scheduled daily @ 15 30 hrs. Its fine.
    CREATE OR REPLACE procedure SCOTT.hrsmail
    is
    v_count number;
    begin
    Select count(*) into v_count from emp;
    if v_count < 1300
    then
    UTL_MAIL.send(sender => '[email protected]',
    recipients => '[email protected]',
    cc => '[email protected]',
    bcc => '[email protected]',
    subject => 'Testing the UTL_MAIL Package',
    message => 'If you get this, UTL_MAIL package
    else
    null; --what you want to do here
    end if ;
    end;
    Sometime the Database is down, so the job is not running.
    How to create a procedure to send a mail if the database is down?
    Pls help me. Its highly appreciated.
    Thanks
    Nihar

    nihar wrote:
    How to create a procedure to send a mail if the database is down?And what if the database is up, but the network down? Or the database up and mail server down? Or mail server undergoing maintenance?
    There are loads of "+What if's+" - and in that respect, playing "+What if database is down..+" in this case does not make any sense. You do not use the database to monitor its own up/down status. You do not rely just on SMTP as notification protocol that the database is down.
    The correct approach would be using something like SNMP as the monitoring protocol. A monitoring system that can process SNMP and perform some basic root cause analysis (e.g. network to the database server down, database server status unknown). And this system supporting notification methods like SMTP, SMS and so on.

  • Additional validation for one of the fields in down payment request

    Hi All ,
      I have a requirement for doing some additional validation for one of the fields in down payment request
    (F-47) .
    I need to validate a field u2018Assignmentu2019 in the down payment request .
    Here, assignment field is mandatory and user has to enter a valid contract number. Then we have to  validate the contract number entered against the tracking number field (BEDNR) in EKPO table corresponding to the vendor.
    If the contract number entered is incorrect, then system displays a message with the valid contract number. The user can modify the assignment field with this contract number. Once the field is modified, system once again validates whether the contract number entered is valid.
    Can you please provide me your valuable inputs how to achieve this.
    Regards,
    Tripti.

    Hi Vishal ,
      Thanks for your reply. I tried searching for  them but couldn't find any.
    Regards,
    Tripti.

  • Backing up the RAC DB when either one of the node is down

    11.2.0.2/Solaris 10 (x86-64bit) For our 2-Node production RAC DB, I had configured RMAN backup from Node1 using Cronjob. Last weekend our Node1 went down. Our SMS notifying system which sends SMS alerts to our Mobiles went down on the weekend as well. Only by Monday Noon we came to know that Node1 is down and that there is no backup for Saturday and Sunday.
    How can i make sure that RMAN backup of the DB will be taken even if either one of the Nodes go down ? My friend suggested IBM TWS scheduler. Can Tivoli Work Scheduler detect a dead RAC Node and fire RMAN backup from the surviving node ?

    I don't know the answer regarding TWS, but if you run the backup from crontab I guess that you don't have any 3rd party tool now.
    I think the easiest solution will be to have the script and crontab job on both servers and decide which one runs the backup.
    For example, the script that is scheduled in the crontab will do:
    1. if $HOSTNAME is node1 run the backup. If $HOSTNAME is node2, check if node1 is up and if not run the backup.
    2. This is more elegant, check the "crsctl status resource" for something and run the backup accordingly. For example, the script will check where SCAN1 VIP is located and this is the node which will run the backup.
    HTH
    Liron

  • Adstrtal.sh: Database connection could not be established. Either the database is down or the APPS credentials supplied are wrong.

    In our R12.1.3 EBS, on RHEL 5.5, 64-bit, we are getting this error -
    adstrtal.sh: Database connection could not be established. Either the database is down or the APPS credentials supplied are wrong.
    However, neither the db is down nor the APPS password is wrong. I am able to connect from sqlplus from application tier. And also from the 10.1.3 home. I have checked similar threads here but none helped. Raised an SR yesterday but that too hasn't helped yet.
    Any ideas?
    Regards,
    Vinod

    please post he output of
    SQL> show parameter sec_case_sensitive_logon
    $ECHO TWO_TASK
    clean FND NODES
    How to Clean Nonexistent Nodes or IP Addresses From FND_NODES [ID 260887.1]
    Check if all services are down on application node, no process should be running.
    shut down db listener and database
    reboot the server
    start or reload listener on DB node and start the database
    Run autoconfig on db tier -- make sure it completes successfully.
    Run autoconfig on all apps tier -- make sure it completes successfully.
    Now, try to start applications on middle tier.
    ApPsMaStI
    sharing is Caring

  • Two rows instead of one in the database when form with 2 data blocks.

    Hi All,
    I have created a form having 2 data blocks
    data block one to populate table A
    data block two to populate table B .
    now when i save the data after pressing the save option ....i find 2 records instead of one in the database table A and Table B.
    it should have been just one per table ....can some one suggest pls
    Sunny

    ok : step by step :
    do you have a trigger created named "ON-INSERT" on
    the block?Many thanks for your support....
    I do not have an ON-insert trigger.....
    I will explain my scenerio.....
    I have 2 data blocks one-below the other in the form
    accessing 2 different tables....
    now on the canvas of this form I have added a
    button....and on that button i have written a trigger
    'when-button-pressed' Trigger....
    it takes the data from the various text_items of the
    form and inserts them into their corrosponding
    tables....and I have added 'COMMIT' at the end this
    trigger
    so when I press this button ....the data 2 records
    get inserted into the 2 tables instead of one each..
    TIA
    Sunnythe above form looks something like this.....
    -------------------------------------------------------------------------------------------------------|
    |------------------------------------------------------------------------------------------------------|
    | |
    | ---data_block1------------------------------------------------------ |
    | | | |
    | | some data entry here for emp1 table | |
    | |__________________________________________ | |
    | |
    | |
    | |
    | ---data_block2------------------------------------------------------ |
    | | | |
    | | some data entry here for dept1 table | |
    | |__________________________________________ | |
    | |
    | button1 |
    |-------------------------------------------------------------------------------------------------------

  • Database Connectivity takes long time if one of the Node is down .. ??

    Hello All,
    Env: 10.2.0.4 on Solaris 10
    I have 2 nodes.
    When Node1 server is down, it takes long time to connec to the database.
    tnsping would give "OK(2050ms)". Below is the tnsalias.
    RAC_test  =
      (DESCRIPTION =
         (ADDRESS = (PROTOCOL = TCP)(HOST=20.268.169.123)(PORT= 1521))
         (ADDRESS = (PROTOCOL = TCP)(HOST=20.268.169.127)(PORT= 1521))
         (LOAD_BALANCE = yes)
              (CONNECT_DATA =
            (SERVICE_NAME = DK.com)
          (FAILOVER_MODE =
            (TYPE = SELECT)
            (METHOD = BASIC)
            (RETRIES = 180)
            (DELAY = 5)
    )I put the trace on sqlnet.ora and found that first it pings to the "20.268.169.123",
    since the Server is down there will not be any reply and this consumes the delay and
    later it would ping "20.268.169.127" and connect to it.
    If i keep "20.268.169.127" above "20.268.169.123" in tnsalias, and keep "LOAD_BALANCE=no",
    it gets connected very fast, as its directly connecting to Node2. In tnsping i get Ok(40ms).
    How do i reduce the connect timing if i use the first step. Why does it take long time for
    Oracle Client to understand that the Node1 Server is down ?
    TIA,
    J J

    I hope the IP's you are using in the TNS are Virtual IP's.
    You must use Virtual IP's / hostnames for the failover to be quick. If Node 1 is not available then then it's (Node 1's) virtual IP would also get assigned to Node 2 hence all client connections are still able to get a response from the Node Virtual IP address without needing to wait for TCP/IP timeouts. This helps clients to get notified immediately that node 1 is unavailable and the connection tries the 2nd ip/host in the connect descriptor.
    Hope this helps.
    - Siba

  • Behavior when one of the instance is down

    Dears,
    We have a reporting server(IBM Cognos BI) using Oracle RAC as the data warehouse.  Cognos uses OCI to communicate with Oracle database(Oracle client is installed with the cognos server machine). Cognos only issues SELECT statements to the database.
    This is our tnsnames.ora file looks like
    ${DB_NAME}=
    (DESCRIPTION=
      (ENABLE=BROKEN)
      (LOAD_BALANCE=off)
      (FAILOVER=on)
      (ADDRESS=(PROTOCOL=tcp)(HOST=${VIP1_IP})(PORT=11521))
      (ADDRESS=(PROTOCOL=tcp)(HOST=${VIP2_IP})(PORT=11521))
      (CONNECT_DATA=
         (SERVICE_NAME=${DB_NAME})
        (FAILOVER_MODE=(TYPE=select)(METHOD=basic))
         (server = dedicated)
    As to TYPE=SELECT, it is said that if the instance on which the select statement is being executed is down, one of the available oracle instances will take over as nothing has happened. The failover is transparent to the application(Cognos). http://www.dba-oracle.com/art_oramag_rac_taf.htm
    But when we manually shutdown an instance , we did not see the result as expected while get the following:
    ORA-25402: transaction must roll back
    Cause: A failure occured while a transaction was active on this connection.
    Action: The client must roll back.
    So, it looks like if application wants to take the advantage of this functionality, the application must be modified to do something related to transaction termination, Right? If so, like a product as Cognos, not a project, it could not do things specific to a particular database. 
    Any suggestion is welcome.
    Ricky
    Thanks,

    HI,
    Please check your configuration as per below link.
    http://www.oracle.com/technetwork/database/features/oci/taf-10-133239.pdf
    FAST CONNECTION FAILOVER
    Fast Connection Failover (FCF) is a packaged implementation of FAN included in Oracle Application Server 10g. In the event of a failure, it will act on the FAN event, clean up stale connections, and pass the notification to the application as a SQL exception. The application can then retry the connection and replay the transaction. It works with JDBC connection pools, and is not dependent on OCI or TAF. Thus, you can use FCF with both the thin JDBC driver as well as the thick (OCI) JDBC driver. Oracle does not recommend the use of both TAF and FCF with the same application.

  • Starting weblogic while the database is down.

    We encountered a problem starting weblogic 8.1.5 while the DB is down.
    The problem is that the connection pool fails to be created. The DS fails to be created as a result.
    When the DB finally comes up, the connection pool does not recover although it does test upon reservation.
    The only way I was able to recover the pool is by undeploying it from the server and then redeploying it. Even then, the DS did not recover; I had to undeploy it and then redeploy to achieve the full recovery.
    I have found a question about this issue from 4 years ago and the suggestion was to create the connection pool with initial size of 0. That won't work for us.
    Any Ideas?
    Oded Blayer
    Invoke Solutions
    [email protected]

    Oded Blayer wrote:
    Thanks for the rapid reply.
    The application is in production stage, and such configuration change requires time to make sure the performance is not harmed.
    The dynamic creation will require even deeper changes.
    Will this issue solved in later versions of weblogic (9, 10)?
    Oded Blayer
    Invoke SolutionsThe connection retry, init-cap=0, and dynamic pool creation
    will probably be the only options for the near future,
    and any new option would be to simply allow the pool to
    come up with no connections, and would then act exactly
    as if you'd set init-cap=0 anyway.
    For your current case, connection retry or init-cap=0
    are probably the most appropriate. Retry will make connections
    soon after the DBMS is available, whereas init-cap=0 will
    populate the pool as application load requires it. The
    performance will only be affected for the first application
    sessions, and will be hurt only by the amount of time it takes
    to make a real DBMS connection. Do not alter the capacity increment.
    Leave it at 1.
    Joe

  • EM displays the database as down

    Hi ,
    Today I migrated a 10g R1 DB to another new server. I changed the server IP to the previous server IP. now the EM shows as the DB is down. but my DB is actually up.
    Is there anything that I have to do ???

    Have you gone into (From Grid Contorl) setup > Agents > Host, then pressed 'Upload Metric data'. Also have you check to see if the host agent is running ?

  • IO discontinued when one of the storage was down, anybody know what's wrong?

    I encountered this issues using the Max OS  X with the ATTO HBA cards and multipath.
    We used two port of the HBA which was connected to one controller of the storage. And we mapped serveral LUNS to the Mac server.That means there were two active paths between the server and the storage.
    Yesterday , one of storage controller which had IO on it was down accidently. And the IO on the volume can't continue until the controller was up.
    Why did this happen?  We had multipath, we want to know why the IO didn't switch to another path.
    Here is log of the server:
    Jul 28 10:00:19 huaweimatoMac-Pro kernel[0]: nstat_lookup_entry failed: 2
    Jul 28 10:01:27 huaweimatoMac-Pro kernel[0]: nstat_lookup_entry failed: 2
    Jul 28 10:07:27 huaweimatoMac-Pro kernel[0]: IOSCSIMultipathedLogicalUnit: Timed out waiting for paths to become available, loops = 0
    Jul 28 10:07:57: --- last message repeated 6 times ---
    Jul 28 10:07:57 huaweimatoMac-Pro kernel[0]: IOSCSIMultipathedLogicalUnit: Timed out waiting for paths to become available, loops = 1
    Jul 28 10:08:27: --- last message repeated 6 times ---
    Jul 28 10:08:27 huaweimatoMac-Pro kernel[0]: IOSCSIMultipathedLogicalUnit: Timed out waiting for paths to become available, loops = 2
    Jul 28 10:08:57: --- last message repeated 6 times ---
    Jul 28 10:08:57 huaweimatoMac-Pro kernel[0]: IOSCSIMultipathedLogicalUnit: Timed out waiting for paths to become available, loops = 3
    Jul 28 10:09:11: --- last message repeated 6 times ---
    Jul 28 10:09:11 huaweimatoMac-Pro kernel[0]: IOSCSIMultipathedLogicalUnit: new active path available, checking, loops = 4
    Jul 28 10:09:41: --- last message repeated 6 times ---
    Jul 28 10:18:43 huaweimatoMac-Pro kernel[0]: hfs: Initializing the journal (joffset 0x50000 sz 0x800000)...
    Jul 28 10:18:43 huaweimatoMac-Pro kernel[0]: ioqueue_depth = 256,   ioscale = 8
    Jul 28 10:19:49 huaweimatoMac-Pro kernel[0]: nstat_lookup_entry failed: 2
    Jul 28 10:21:53 huaweimatoMac-Pro kernel[0]: HFS: Low Disk: freeblks: 38287, warninglimit: 38400

    Make sure all apps are fully closed:
    - double tap your home button
    - hold your finger on an app
    - when the apps start to wiggle press the red "-" button on all the apps
    Now do a hard reset:
    - hold the home button and the sleeps/wake button down at the same time
    - the phone will shut off then when the apple logo ppears release both buttons
    If by then your contacts have not been restored you would need to try and retreive them from a back up on itunes/icloud that was made when the contacts were on the phone. You could even stream them from the cloud if they were synced to the cloud. The link beow will show you how to restore from an itunes/icloud back up:
    http://support.apple.com/kb/ht1766
    Also to prevent this issue which i have seen alot when users are still running iOS 5 it would be best to update the iOS on your phone. To do this follow the link below:
    http://support.apple.com/kb/ht4623
    Hope this helps

  • How Can I Open More Pictures To Work On Them As It Was Before Because Now I Can Open Just One Picture The Others Go Down If I Need One I Have To Click It To Come To The Screen

    Hi My Problem Is Hen I Tried To Open More Than One Pictures To Work On It Open The First Picture The Second One Goes Down On The panel So I Cannot Have All Together On The Screen As It Was Before, I Uninstalled Adobe Photoshop Element 10 And Reinstall It Again But No Luck I Do Have The Same Problem Please Can I Have Help, Thank You

    Hi [email protected],
    First go to your Preferences for Photoshop Elements 9. In Windows go up to the Edit menu and choose Preferences. On a Mac go up to the Photoshop Elements menu and choose Preferences.
    From Preferences choose General…. Look for where it says Allow Floating Documents in Full Edit Mode and check the box in front of it if it's not already checked. Click OK.
    Now go up to the Window menu and choose Images>Float All in Windows.
    That should show all your open photos as separate windows.
    Regards,
    Sarika

  • 4 ipods sharing music on ONE computer GET one UP the others GO DOWN

    We run windows and have 4 ipods connected to the computer all with different log on accounts
    We share all our music with each other
    Problem
    when we search for a song and click on it (so we can get it in our individual itunes) seems when you log on to a different ipod theirs is all messed up with a exclamation point on the song just previously shared.
    Also, I have noticed when I search for a song I find we have up to 6 and at times 10 copies of the same song. Is there a easy way to clean up this mess and still keep all 4 ipods running?

    Well, the fiirst option wasn't working because nothing will sync unless iTunes is running. Since iTunes no longer warns you before it starts syncing, when you move your iPod to a different computer, that means any kind of auto-syncing of my music was not an option.
    I also got an error message each time I synced with my iMac after removing iPhoto from it (could not find a compatible photo program), which was already getting annoying. Removing Address Book from the iMac did allow me to sync contacts with the PowerBook and not lose them when syncing to the iMac.
    Still, this was only 1/3 of what I was trying to accomplish, and I was getting the feeling that I was just fighting against the way this was supposed to work. So I switched to the second option.
    This isn't too bad with the help of Sync Pro X, which I already have installed on my PowerBook for daily backups. What I'm doing now is that whenever I update my photos or contacts on the PowerBook, I run a Sync Pro X script that connects to my iMac and copies my changes in those two programs. Then I can do a full, automatic sync of my iPod to my iMac. Under this plan, I'll never try to sync the iPod to the PowerBook -- it's just a one-computer, one-iPod situation.
    I've been following this plan for a week, and it's pretty smooth. It's not exactly what I had in mind, but I've come to think that it's the best approach given how the iPod is designed to work.
    I hope this little experiment helps someone, if anyone is trying to do what I was doing.
    Cheers,
    -Arlo

Maybe you are looking for