Incompatible archive logs dest and dest_n

Hello all,
My problem is the following:
I have put my database in the archivelog mode.
But I can't fill in my local directory for the arc files.
The message says that archive_log_dest and archive_lod_dest_duplex are incompatible with mine archive_log_dest_1.
Still I believe that the former archive_log_dest are not in use.
What should I do?
Thanks in advance,
Hans Puts

Hello Hans,
I hope I understood you're question right and you've switched on the
archive-logging online, so one thing you can do is the following:
stop archive-logging,
shutdown db,
edit the init.ora file and activate the archive-logging there and
set also the destination-directories there (don't use system variables!),
start db again
Normally archive-logging should work then.
Greetings Tobias

Similar Messages

  • RMAN archive logs backup and restore in RAC environment.

    Hello All,
    I have a 8 node RAC using ASM production database and a single node standby database using ASM. I have archive log gap on the standby database and the archive logs on the primary database is backed up and removed.
    I need to resync the standby database using RMAN backup/restore.
    Could you please help me on how to backup the specific archive logs file (all threads) from production and restore/recover the standby database. Any specific RMAN script or Metalink ID will help alot.
    Details:
    Source: 8 node RAC, ASM, Linux,
    target: Non-RAC, ASM, Linux.
    Regards,

    You can try LIKE clause of BACKUP ARCHIVELOG such as:
    RMAN> backup archivelog like '%XE%2010%20%';
    Starting backup at 20-AUG-10
    using channel ORA_DISK_1
    channel ORA_DISK_1: starting archive log backupset
    channel ORA_DISK_1: specifying archive log(s) in backup set
    input archive log thread=1 sequence=58 recid=19 stamp=727519553
    input archive log thread=1 sequence=59 recid=20 stamp=727544121
    input archive log thread=1 sequence=60 recid=21 stamp=727544127
    channel ORA_DISK_1: starting piece 1 at 20-AUG-10
    channel ORA_DISK_1: finished piece 1 at 20-AUG-10
    piece handle=C:\ORACLEXE\APP\ORACLE\FLASH_RECOVERY_AREA\XE\BACKUPSET\2010_08_20\
    O1_MF_ANNNN_TAG20100820T154401_66X1O353_.BKP tag=TAG20100820T154401 comment=NONE
    channel ORA_DISK_1: backup set complete, elapsed time: 00:00:04
    Finished backup at 20-AUG-10or you can try SEQUENCE clause:
    RMAN> backup archivelog sequence between 58 and 60 thread 1;
    Starting backup at 20-AUG-10
    using channel ORA_DISK_1
    channel ORA_DISK_1: starting archive log backupset
    channel ORA_DISK_1: specifying archive log(s) in backup set
    input archive log thread=1 sequence=58 recid=19 stamp=727519553
    input archive log thread=1 sequence=59 recid=20 stamp=727544121
    input archive log thread=1 sequence=60 recid=21 stamp=727544127
    channel ORA_DISK_1: starting piece 1 at 20-AUG-10
    channel ORA_DISK_1: finished piece 1 at 20-AUG-10
    piece handle=C:\ORACLEXE\APP\ORACLE\FLASH_RECOVERY_AREA\XE\BACKUPSET\2010_08_20\
    O1_MF_ANNNN_TAG20100820T154911_66X1YRJL_.BKP tag=TAG20100820T154911 comment=NONE
    channel ORA_DISK_1: backup set complete, elapsed time: 00:00:04
    Finished backup at 20-AUG-10or
    RMAN> backup archivelog sequence between 58 and 60;
    Starting backup at 20-AUG-10
    using channel ORA_DISK_1
    channel ORA_DISK_1: starting archive log backupset
    channel ORA_DISK_1: specifying archive log(s) in backup set
    input archive log thread=1 sequence=58 recid=19 stamp=727519553
    input archive log thread=1 sequence=59 recid=20 stamp=727544121
    input archive log thread=1 sequence=60 recid=21 stamp=727544127
    channel ORA_DISK_1: starting piece 1 at 20-AUG-10
    channel ORA_DISK_1: finished piece 1 at 20-AUG-10
    piece handle=C:\ORACLEXE\APP\ORACLE\FLASH_RECOVERY_AREA\XE\BACKUPSET\2010_08_20\
    O1_MF_ANNNN_TAG20100820T154748_66X1W5Z0_.BKP tag=TAG20100820T154748 comment=NONE
    channel ORA_DISK_1: backup set complete, elapsed time: 00:00:04
    Finished backup at 20-AUG-10See http://download.oracle.com/docs/cd/B19306_01/backup.102/b14194/rcmsynta008.htm#i78730

  • Clone database with no archive log mode and downtime with new file structure

    We need to clone database with following business requirements :
    No downtime at all can be afford
    databases are in no archive log mode
    data-file locations need to be change to new mount points while cloning for space management factor
    Please suggest the best possible methods for same ?

    Can you post your version of oracle in 4 digits and OS for better understandig
    I dont think you can move the No Archive database to a new location without shutting down.
    You just want to move them to the new mount points is it?
    1. How big are your datafiles are?
    2.How do you backup this DB?
    When you move the datafilles then it becomes inconsistent and you need redo logs and probably archive logs to be there to make the file consistent or else it would be prompting for recovery. The chances are your redolog will be overwritten and hence you wont be able to recover.
    You need downtime.

  • How to calculate storage space for archive log files and database backups?

    Hi all,
    I have a 1.8 terabyte Oracle 9i database and need to plan for how much additional disk space I will need to perform nightly backups and for archivelog files. Is there a script or formula available that can help me estimate how much required disk space I will need to hold a days worth of archived logs as well as a nightly export dump file and a full hot RMAN backup on disk?
    Thanks!

    I'm not sure how to estimate the size of your backups, especially if you use incrementals. However, the space required for archive logs will be equal to the amount of REDO your DB generates. I would count the number of log switches per day with a query like the following:
    select trunc(first_time), count(*)
    from v$log_history
    group by trunc(first_time)
    I would then take the average and multiply this count by the size of your redo log files (assuming they are all the same size).

  • Why backing up archive logs? (and more...)

    Hi all,
    If I schedule a weekly backup every Saturday:
    BACKUP DATABASE PLUS ARCHIVELOG;
    1- Does the command backup all the archivelogs generated during the week (and what for as far as I have the full backup) or only the archivelogs generated during the backup execution time?
    2- If I use a daily incremental backup with a weekly merge strategy , does the weekly archivelogs needed (or only one day) in case of recovery?
    3- The incremental Merge operation only uses the incremental backupset and the previous full backup to merge into datafile copies. No archivelogs are used to do this merge operation?
    4- Does the following strategy meaningfull regarding the archivelogs (incremental and merge backup every day)?
    RUN {
    BACKUP INCREMENTAL LEVEL 1 FOR RECOVER OF COPY WITH TAG 'INCR' DATABASE;
    RECOVER COPY OF DATABASE WITH TAG 'INCR';
    DELETE NOPROMPT OBSOLETE;
    BACKUP ARCHIVELOG ALL DELETE INPUT;
    Thanks,
    Vince

    BACKUP DATABASE PLUS ARCHIVELOG;
    1- Does the command backup all the archivelogs generated during the week (and what for as far as I have the full backup) or only the archivelogs generated during the backup execution time?
    The above command will first generate a log switch then backup the archivelogs which have not been backed up so far then backup the database (full) and then switch the log file again and backup the remainining archivelogs.
    2- If I use a daily incremental backup with a weekly merge strategy , does the weekly archivelogs needed (or only one day) in case of recovery?
    RMAN always take incremental over archivelogs so if the incremental backup is available then archivelogs are not required BUT let say all the incrementals are taken as 2 am and you want to recover the database upto 4pm so next day incremental wont work then RMAN will use all incrementals upto the day of failure and then use the archivelogs for remaining recovery. In short you should decide the recovery window of your database means how far back you wanna go in case of incomplete recovery and you MUST keep all the archivelogs for that period of time.
    3- The incremental Merge operation only uses the incremental backupset and the previous full backup to merge into datafile copies. No archivelogs are used to do this merge operation?
    I hope answer number 2 will help here too.
    4- Does the following strategy meaningfull regarding the archivelogs (incremental and merge backup every day)?
    RUN {
    BACKUP INCREMENTAL LEVEL 1 FOR RECOVER OF COPY WITH TAG 'INCR' DATABASE;
    RECOVER COPY OF DATABASE WITH TAG 'INCR';
    DELETE NOPROMPT OBSOLETE;
    BACKUP ARCHIVELOG ALL DELETE INPUT;
    If this is the only backups you are having so it means your recovery window is 1 day because your backup copies are a day older than your current database and you can't use these copies to restore more than a day back. BUT you still want to backup archivelogs for last day so that in case of failure you can use restore upto a specific time and you dont want the archilogs more than a day older, those are useless because there is no prior backup available where you can apply those.
    Daljit Singh

  • Changing the location of archive log from flash recovery area PLZ HELP!!!

    Hi All,
    My archive log is being stored in flash memory area which got full and the production server went down.
    alert log file details.....
    ORA-19809: limit exceeded for recovery files
    ORA-19804: cannot reclaim 43432960 bytes disk space from 2147483648 limit
    *** 2010-04-25 14:22:49.777 62692 kcrr.c
    ARCH: Error 19809 Creating archive log file to
    '/oracle/product/10.2.0/flash_rec
    overy_area/EDWREP/archivelog/2010_04_25/o1_mf_1_232_%u_.arc'
    *** 2010-04-25 14:22:49.777 60970 kcrr.c
    kcrrfail: dest:10 err:19809 force:0 blast:1I removed the files and started the database,
    Can someone kindly tell me as to how to avoid this problem in future by keeping archive log destination in flash recovery area.
    I want to change the location of archive log files, can someone please guide me as to hiow to do that
    I changed the size of flash recovery area for the time being, but i am afraid it will be full again!!
    SQL> select * from v$flash_recovery_area_usage;
    FILE_TYPE    PERCENT_SPACE_USED PERCENT_SPACE_RECLAIMABLE NUMBER_OF_FILES
    CONTROLFILE                   0                         0               0
    ONLINELOG                     0                         0               0
    ARCHIVELOG                99.44                         0              57
    BACKUPPIECE                   0                         0               0
    IMAGECOPY                     0                         0               0
    FLASHBACKLOG                  0                         0               0
    6 rows selected.
    SQL> alter system set DB_RECOVERY_FILE_DEST_SIZE = 4G ;
    System altered.
    SQL> select * from v$flash_recovery_area_usage;
    FILE_TYPE    PERCENT_SPACE_USED PERCENT_SPACE_RECLAIMABLE NUMBER_OF_FILES
    CONTROLFILE                   0                         0               0
    ONLINELOG                     0                         0               0
    ARCHIVELOG                49.72                         0              57
    BACKUPPIECE                   0                         0               0
    IMAGECOPY                     0                         0               0
    FLASHBACKLOG                  0                         0               0
    6 rows selected.regards,
    Edited by: user10243788 on Apr 25, 2010 6:12 AM

    user10243788 wrote:
    Hi All,
    My archive log is being stored in flash memory area which got full and the production server went down.
    alert log file details.....
    ORA-19809: limit exceeded for recovery files
    ORA-19804: cannot reclaim 43432960 bytes disk space from 2147483648 limit
    *** 2010-04-25 14:22:49.777 62692 kcrr.c
    ARCH: Error 19809 Creating archive log file to
    '/oracle/product/10.2.0/flash_rec
    overy_area/EDWREP/archivelog/2010_04_25/o1_mf_1_232_%u_.arc'
    *** 2010-04-25 14:22:49.777 60970 kcrr.c
    kcrrfail: dest:10 err:19809 force:0 blast:1I removed the files and started the database,
    Can someone kindly tell me as to how to avoid this problem in future by keeping archive log destination in flash recovery area.
    I want to change the location of archive log files, can someone please guide me as to hiow to do that
    I changed the size of flash recovery area for the time being, but i am afraid it will be full again!!
    SQL> select * from v$flash_recovery_area_usage;
    FILE_TYPE    PERCENT_SPACE_USED PERCENT_SPACE_RECLAIMABLE NUMBER_OF_FILES
    CONTROLFILE                   0                         0               0
    ONLINELOG                     0                         0               0
    ARCHIVELOG                99.44                         0              57
    BACKUPPIECE                   0                         0               0
    IMAGECOPY                     0                         0               0
    FLASHBACKLOG                  0                         0               0
    6 rows selected.
    SQL> alter system set DB_RECOVERY_FILE_DEST_SIZE = 4G ;
    System altered.
    SQL> select * from v$flash_recovery_area_usage;
    FILE_TYPE    PERCENT_SPACE_USED PERCENT_SPACE_RECLAIMABLE NUMBER_OF_FILES
    CONTROLFILE                   0                         0               0
    ONLINELOG                     0                         0               0
    ARCHIVELOG                49.72                         0              57
    BACKUPPIECE                   0                         0               0
    IMAGECOPY                     0                         0               0
    FLASHBACKLOG                  0                         0               0
    6 rows selected.regards,
    Edited by: user10243788 on Apr 25, 2010 6:12 AMPointing the archive log dest (and/or the FRA) to a new location, or enlarging them, will do no good if you are not performing regular housekeeping on the archivelogs. You will just keep knocking down the same problem over and over.
    If you simply delete the archivelogs at the OS level, the database will never know about it and it will continue to think the destination is full, based on records kept in the control file.
    For regular housekeeping, you need to be doing something similar to this in rman:
    run {
      backup archivelog all not backed up 1 times tag='bkup_vlnxora1_arch';
      delete noprompt archivelog all backed up 1 times to device type disk;
    run {
    delete noprompt obsolete;
    crosscheck archivelog all;
    delete noprompt expired archivelog all;

  • RAC online and archive logs question

    Hello All,
    I setup a RAC database instances prod1 and prod2 (10.2.0.4). Datafiles and onlinelogs are on ASM.
    Does these results look good queried from two instances. I am kind of concerned about the Group3 that has the same name for both the members.
    Also archived logs are going to the ASM, is this a good practice. I was reading Oracle RMAN book and it mentioned archived logs go to local disk.
    Is it possible to archive to local disk for online that are on ASM? Please advice. Early reply appreciated.. Thanks San~
    PROD1 Instance
    SQL> select member from v$logfile;
    MEMBER
    +DATA/prod/onlinelog/group_2.264.706892209
    +FLASH/prod/onlinelog/group_2.259.706892211
    +DATA/prod/onlinelog/group_1.261.706892209
    +FLASH/prod/onlinelog/group_1.260.706892209
    +DATA/prod/onlinelog/group_3.258.706892235
    +FLASH/prod/onlinelog/group_3.258.706892235
    +DATA/prod/onlinelog/group_4.256.706892237
    +FLASH/prod/onlinelog/group_4.257.706892237
    8 rows selected.
    PROD2 Instance
    SQL> select member from v$logfile;
    MEMBER
    +DATA/prod/onlinelog/group_2.264.706892209
    +FLASH/prod/onlinelog/group_2.259.706892211
    +DATA/prod/onlinelog/group_1.261.706892209
    +FLASH/prod/onlinelog/group_1.260.706892209
    +DATA/prod/onlinelog/group_3.258.706892235
    +FLASH/prod/onlinelog/group_3.258.706892235
    +DATA/prod/onlinelog/group_4.256.706892237
    +FLASH/prod/onlinelog/group_4.257.706892237
    8 rows selected.
    ===
    SQL> archive log list
    Database log mode Archive Mode
    Automatic archival Enabled
    Archive destination USE_DB_RECOVERY_FILE_DEST
    Oldest online log sequence 3
    Next log sequence to archive 4
    Current log sequence 4
    ====
    Thanks
    San

    Hi San,
    sannidhi wrote:
    Also archived logs are going to the ASM, is this a good practice. I was reading Oracle RMAN book and it mentioned archived logs go to local disk.
    Is it possible to archive to local disk for online that are on ASM? Please advice. Early reply appreciated.. Thanks San~
    It is recommend to store archived log files on ASM and on Shared disk, check your archive log format which suppose to represent uniqueness across all instances.
    Yes, technically it is possible to archive to local disk, but not recommended as if you loose local disk there will be gaps in the archived log files and also it increases the administration.
    Regards,
    Thota

  • When to delete Archive Log files ? and is safe to delete ?

    Hi all,
    I have a question, my DB is running in Archive Log mode and is growing day by day, so should I delete archive log files and where i find that my which files are old and how to delete theses files, for info I also take full RMAN and export backup regularly .
    abbas

    Well, if you're not comfortable with Backup and Recovery principles, i'd advise you to first read TFM: Backup and Recovery Concept(click) for you data safety's sake.
    Anyway, just to give you a glance at what could happen, immagine suche a situation where your database is made of 5 different datafiles (whatever they are - SYSTEM, SYSAUX, UNDO, ...), controlfiles and online redo logs. If, for example your backup is done in a timeline such as:T0 : Backup datafile F1 and F2
    T1 : Backup datafile F2 and controlfile
    T2 : Backup datafile F3, F4
    T3 : Backup datafile F5
    T0' : Backup datafile F1 and F2
    T1' : Backup datafile F2 and controlfile
    T2' : Backup datafile F3, F4
    T3' : Backup datafile F5
    ...If your server crashes (for example is thrown though the 5th floor window..) in order to make a full recovery, you could proceed multiple ways.
    . Restore files from Tx' backups. Then you'd need every archived redo logs created from T0' to now in order to be able to do a full recovery.
    . Restore files from backups; T3, T0', T2' and rectreate controlfile. Then you'd need every archived redo log created from T3 (inclusive).
    . Restore Tx backup, ...
    . Restore ...
    and so on.
    What's important is that you're at least able to resynch a full-set backup. What I mean is: if you restored "T3, T0', T2' and rectreate controlfile", in order to open the database you'll absolutely need every archived redo log created between T3 and T2' inclusive. If one is missions, you won't be able to open it.
    That is the principle. Read the concept book I linked to get more infos.
    Regards,
    Yoann.

  • How to delete the data in archived log files

    hi
    how can i delete the enteries in archived log files. and what is the disadvantage of deleting archived log enteries.

    There is no documented way to delete data stored in archived log files: you can only remove the archived log files if needed.

  • How to remove archive log files using Enterprise Manger

    Hi,
    I took backup of the archive log files and i deleted those files at OS level but it is not reflecting in the Enterprise Manager it is showing the same. Can any one help in how to delete the archive log files by using Enterprise Manger.
    version 10.2.0.1.0
    any help is appreciated.
    Regards,
    Ashraf

    I have the same problem. I have 10gr1 and 10gr2 on different servers. On 10gr1, under administration/manage current backup/image, I can see archivelog files, and can mange them. But on 10gr2, the image doesn't show archivelog files. I can only remove them using rman. I wonder if it's because the different design on 10gr2 or I missed something when setup the backup?

  • How to determine which archive logs are needed in flashback.

    Hi,
    Let's assume I have archive logs 1,2,3,4, then a "backup database plus archivelogs" in RMAN, and then archive logs 5+6. If I want to flashback my database to a point immediately after the backup, how do I determine which archive logs are needed?
    I would assume I'd only need archive logs 5 and/or 6 since I did a full backup plus archivelogs and the database would have been checkpointed at that time. I'd also assume archive logs 1,2,3,4 would be obsolete as they would have been flushed to the datafiles in the checkpoint.
    Are my assumptions correct? If not what queries can I run to determine what files are needed for a flashback using the latest checkpointed datafiles?
    Thanks.

    Thanks for the explanation, let me be more specific with my problem.
    I am trying to do a flashback on a failed primary database, the only reason why I want to do a flashback is because dataguard uses the flashback command to try and synchronize the failed database. Specifically dataguard is trying to run:
    FLASHBACK DATABASE TO SCN 865984
    But it fails, if I run it manually then I get:
    SQL> FLASHBACK DATABASE TO SCN 865984;
    FLASHBACK DATABASE TO SCN 865984
    ERROR at line 1:
    ORA-38754: FLASHBACK DATABASE not started; required redo log is not available
    ORA-38761: redo log sequence 5 in thread 1, incarnation 3 could not be accessed
    Looking at the last checkpoint I see:
    CHECKPOINT_CHANGE#
    865857
    Also looking at the archive logs:
    RECID STAMP THREAD# SEQUENCE# FIRST_CHANGE# FIRST_TIM NEXT_CHANGE# RESETLOGS_CHANGE# RESETLOGS
    25 766838550 1 1 863888 10-NOV-11 863892 863888 10-NOV-11
    26 766838867 1 2 863892 10-NOV-11 864133 863888 10-NOV-11
    27 766839225 1 3 864133 10-NOV-11 864289 863888 10-NOV-11
    28 766839340 1 4 864289 10-NOV-11 864336 863888 10-NOV-11
    29 766840698 1 5 864336 10-NOV-11 865640 863888 10-NOV-11
    30 766841128 1 6 865640 10-NOV-11 865833 863888 10-NOV-11
    31 766841168 1 7 865833 10-NOV-11 865857 863888 10-NOV-11
    How can I determine what archive logs are needed by a flashback command? I deleted any archive logs with a SCN less than the checkpoint #, I can restore them from backup but I am trying to figure out how to query what is required for a flashback. Maybe this coincides with the point that flashbacks have nothing to do with the backups of datafiles or the checkpoints?

  • Archive log backupset deletion

    Hi all,
    DB version is 10.2.0.4 Have recently changed the retention policy from backup from 2 to 1. After changing the redundancy to 1 Datafile backup sets are getting deleted after backup and all archivelog backupsets remain in the backup location. How do i delete those archivelog backupsets during backup.
    thanks,
    baskar.l

    baskar.l wrote:
    Hi all,
    DB version is 10.2.0.4 Have recently changed the retention policy from backup from 2 to 1. After changing the redundancy to 1 Datafile backup sets are getting deleted after backup and all archivelog backupsets remain in the backup location. How do i delete those archivelog backupsets during backup.
    thanks,
    baskar.lSince these are backups of archivelogs, their final determination of 'obsolete' has to be measured against the backupsets of the datafiles. Are you also taking regular backups of the datafiles?
    for instance, you set retention policy to redundancy of 1. You take a full, Incremental 0 backup. Then you take an incremental level 1 backup. Then you take another incremental level 1 backup. Then you take a backup of your archivelogs. THen you take another backup of archive logs. And another. At this point, nothing is obsolete, because it is all still required to perform a recovery from that one and only full backup.

  • Archive log mode doubt

    Hi All,
    We have 10g R2 database running on RHEL 4 OS. In order to set the archive log mode for the database, if we set only log_archive_dest_n parameter is only enough or we have to set db_recovery_file_dest_size and db_recovery_file_dest also? i know that these 2 parameters are required to configure the flash_recovery_area. my doubt is whether, do we need to set flash_recovery_area parameters and log_archive_dest_n parameters to take backup of databases or set only the log_archive_dest_n parameter and enable archive log mode and take backup using RMAN?
    Thanks,
    Praveen.

    Hi,
    To use RMAN to take a hot backup you need only set LOG_ARCHIVE_DEST_n and place the database in archive log mode. RMAN can be used to backup changed data even after a nologging operation, and you can move away from archivelog mode.
    http://www.stanford.edu/dept/itss/docs/oracle/10g/server.101/b10726/configbp.htm
    There is also the new "block change tracking" B&R that does not need archivelog mode:
    http://www.dba-oracle.com/t_block_change_tracking.htm

  • Archive log mode required for async. CDC?

    Hello there,
    I have tried to set up asyn. HotLog CDC on my 10g database but after creating a publisher I get the following error Message when creating a change set with the publisher:
    BEGIN
    DBMS_CDC_PUBLISH.CREATE_CHANGE_SET(
    change_set_name => 'MY_CHANGESET',
    description => 'First Change set',
    change_source_name => 'HOTLOG_SOURCE',
    stop_on_ddl => 'y',
    begin_date => sysdate,
    end_date => sysdate+20);
    END;
    following Error-Message:
    BEGIN
    ERROR at line 1:
    ORA-00258: manual archiving in NOARCHIVELOG mode must identify log
    ORA-06512: at "SYS.DBMS_CAPTURE_ADM_INTERNAL", line 115
    ORA-06512: at line 1
    ORA-06512: at "SYS.DBMS_CDC_PUBLISH", line 57
    ORA-06512: at line 2
    Does this mean that I have to activate the ARCHIVE LOG-mode, and if yes, how do I do this with the Enterprise Manager?
    I already set the LOG_ARCHIVE_START initparam to "true".
    Klaus

    To place a database in archivelog mode you must:
    Be sure that the log_archive_start and log_archive_dest
    parameters are set in the init file.
    Shut the instance down.
    startup mount;
    alter database archivelog;
    alter database open;

  • Archive log / nologging/ direct path insert

    Could you please confirm if following are true or correct me if my understanding is wrong:
    1 ) Archive log mode and LOGGING is needed to deal with media recovery; it was not needed for instance recovery.
    2) IF insert is in NO APPEND mode , redo is generated even if table is in nologging mode and database is in noachive log mode. This redo is needed for instance recovery.
    3) Direct path insert skips undo generation and may skip redo generation if the object is in nologging mode.
    Thanks.
    In case if it is relevant , I am using Oracle 11.2.0.3.

    1) Yes, Archive logs are needed for media recovery.
    2 and 3) Even if the table is in nologging mode , it generates little bit of redo for index maintenance and dictionary data. Upon a restart from a failure - Oracle will read the online redo logs and replay any transaction it finds in there. That is the "roll forward" bit. The binary redo information is used to replay everything that did not get written to the datafiles. This replay included regenerating the UNDO information (UNDO is protected by redo).
    After the redo has been applied, the database is typically available for use now - and the rollback phase begins. For any transaction that was being processed when the instance failed - we need to undo its changes, roll it back. We do that by processing the undo for all uncommitted transactions.
    The database is now fully recovered.
    Also read he following link
    http://docs.oracle.com/cd/B19306_01/server.102/b14220/startup.htm
    http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:5280714813869

Maybe you are looking for