13 Replies Latest reply: Dec 30, 2013 3:53 AM by Aman.... RSS

redo log file thread

monoj Newbie
Currently Being Moderated

Dear All,

 

 

In am using oracle 10.2.0.4.0 in RHEL 4.

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

 

In my current set we had two node RAC (tims1,tims2) db_name (tims).

 

One of the node has failed(hard disk),Failed node information has been remove from OCR(all the cluster component like...ons,vip,listener,instance).

 

My Question IS:After Removing All the Cluster Component does it remove redo log files associate with the failed instance.

 

Redo logfile Before Node Delete:

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

 

SQL> select member from v$logfile;

 

MEMBER

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

/database/tims/redo02.log

/database/tims/redo01.log

/database/tims/redo03.log

/database/tims/redo04.log



Redo Logfile After Node Delete

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



SQL> select member from v$logfile;

 

MEMBER

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

/database/tims/redo02.log

/database/tims/redo01.log

/database/tims/redo03.log

/database/tims/redo04.log

 

 

It means redo logfile has not been removed.

 

Please help me what are the redo thread i will have to remove.

 

SQL> select group#,sequence#,thread# from v$log;

 

    GROUP#  SEQUENCE#    THREAD#

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

         1      28868          1

         2      28867          1

         3      29327          2

         4      29328          2

 

 

 

Thanks &Regards

Monoj Das

  • 1. Re: redo log file thread
    sb92075 Guru
    Currently Being Moderated

    Is the failed node still down & unavailable?

     

    I suspect that posted SQL & results are only from the surviving node.

     

    Why do you think that you need to do something with the REDO log files?

  • 2. Re: redo log file thread
    monoj Newbie
    Currently Being Moderated

    @

     

    Thanks For you Reply.....Yes you are correct,the sql and result is from surviving nodes.

     

    Yes failed node is down and unavailable and all the cluster related component has been remove for the failed node.

     

    I'm not sure about the deletion of REDO log from the surviving node.Does it require to delete the REDO log?

     

     

    Thanks and Regards

    Monoj

  • 3. Re: redo log file thread
    sb92075 Guru
    Currently Being Moderated

    Why do you think that you need to do something with the REDO log files?

  • 4. Re: redo log file thread
    Aman.... Oracle ACE
    Currently Being Moderated

    monoj wrote:

     

    @

     

    Thanks For you Reply.....Yes you are correct,the sql and result is from surviving nodes.

     

    Yes failed node is down and unavailable and all the cluster related component has been remove for the failed node.

     

    I'm not sure about the deletion of REDO log from the surviving node.Does it require to delete the REDO log?

     

     

    Thanks and Regards

    Monoj

    Monoj,

     

    The redo log files are stored on the shared storage. A node's deletion won't be calling for the files to be automatically deleted. Did you try to add the node back and enable the thread again using these existing redo log files of the failed thread? What's the status of the thread and of the log groups?

     

    Aman....

  • 5. Re: redo log file thread
    Hemant K Chitale Oracle ACE
    Currently Being Moderated

    >One of the node has failed(hard disk),

    That would mean that the failed disk was a *local* disk, not on shared storage.

     

    >Failed node information has been remove from OCR(all the cluster component like...ons,vip,listener,instance).

    Why ?

     

    >My Question IS:After Removing All the Cluster Component does it remove redo log files associate with the failed instance

    No.  Also remember that Redo Log files are on *shared* storage.

     

    Why do you think / want to remove the redo log files ?

     

     

    Hemant K Chitale

  • 6. Re: redo log file thread
    monoj Newbie
    Currently Being Moderated

    Dear All

     

    SQL> select group#,sequence#,thread# from v$log;

     

        GROUP#  SEQUENCE#    THREAD#

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

             1      28868          1

             2      28867          1

             3      29327          2

             4      29328          2

     

     

    According to my above query, thread# for group 1,2 is showing as 1 but the thread has been remove from cluster is there any use of those redo log.

     

     

    Thanks and Regards

    Monoj Das

  • 7. Re: redo log file thread
    Aman.... Oracle ACE
    Currently Being Moderated

    monoj wrote:

     

    Dear All

     

    SQL> select group#,sequence#,thread# from v$log;

     

        GROUP#  SEQUENCE#    THREAD#

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

             1      28868          1

             2      28867          1

             3      29327          2

             4      29328          2

     

     

    According to my above query, thread# for group 1,2 is showing as 1 but the thread has been remove from cluster is there any use of those redo log.

     

     

    Thanks and Regards

    Monoj Das

    Please leave your theories aside and read the replies given in the thread. The logs are going to stay as they are (should be) on the shared storage. Why you are hard-bent to remove these log files anyways? Are you not willing to add the other instance? RAC db with just one instance is really not of much use.

     

    Aman....

  • 8. Re: redo log file thread
    jgarry Guru
    Currently Being Moderated

    It's pretty useful compared to the node with the bad disk.

  • 9. Re: redo log file thread
    Aman.... Oracle ACE
    Currently Being Moderated

    jgarry wrote:

     

    It's pretty useful compared to the node with the bad disk.

    Hehehe yeah but as it is very commonly said now-"disk is cheap" and if I may add to it-"RAC is costly" than having a node working but paying for it's licensing cost doesn't sound like a very good thing to do ;-) !!

     

    Regards

    Aman....

  • 10. Re: redo log file thread
    jgarry Guru
    Currently Being Moderated

    I've seen it where most of the stuff is on SAN, but some local disk came with the machine, "so why waste it."  If you don't have another one handy, it might not be so cheap.

  • 11. Re: redo log file thread
    monoj Newbie
    Currently Being Moderated

    Yes ,I already added another instance to my existing cluster.

     

    This time instance has been added with new thread number i.e 3.

  • 12. Re: redo log file thread
    monoj Newbie
    Currently Being Moderated

    Dear Aman Sir,

     

    Yes ,I already added another instance to my existing cluster.

     

    This time instance has been added with new thread number i.e 3.

     

    SQL> select group#,sequence#,thread# from v$log where thread#=3;

     

        GROUP#  SEQUENCE#    THREAD#

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

             5      28891          3

             6      28890          3

  • 13. Re: redo log file thread
    Aman.... Oracle ACE
    Currently Being Moderated

    Great

     

    Aman....

Legend

  • Correct Answers - 10 points
  • Helpful Answers - 5 points