Updation of controlfile - Page 2
DBAsupport.com Forums - Powered by vBulletin
Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 22

Thread: Updation of controlfile

  1. #11
    Join Date
    Aug 2002
    Location
    Bangalore, India
    Posts
    405
    Originally posted by jmodic
    Nagarjuna,

    However I'm sure of one thing: It is certainly not session's server process. Those processes never write anything to the database, all of their I/O activities are READs and READs only.
    nah.. Server processes do coalesc when you use the 'alter tablespace tsname coalesce;' command.
    -nagarjuna

  2. #12
    Join Date
    Feb 2002
    Posts
    267
    ....and server proscess is the one which writes from DBbuffer cache to the redo log buffer

  3. #13
    Join Date
    Dec 2000
    Location
    Ljubljana, Slovenia
    Posts
    4,439
    Originally posted by nagarjuna
    nah.. Server processes do coalesc when you use the 'alter tablespace tsname coalesce;' command.
    Where did you hear that????? Server proces *can never* write to the database file. Never! It can only read from database file. It would be a shocking news to me if you could prove me wrong on that.

    Why would be any different with ALTER TABLESPACE COALESCE command? How does it differ with any other CREATE, ALTER, etc commands? No matter what command is executed, server process can only *read* from datafiles. All modifications that must be performed by server processare are executed in the buffer cache, and it is DBWR that performs any datafile writes associated with that.

    Realy, can you provide any explanation/URL regarding what is so special about ALTER TABLESPACE COALESCE in this context?
    Jurij Modic
    ASCII a stupid question, get a stupid ANSI
    24 hours in a day .... 24 beer in a case .... coincidence?

  4. #14
    Join Date
    Apr 2001
    Location
    Czechia
    Posts
    712
    Originally posted by Sonia
    ....and server proscess is the one which writes from DBbuffer cache to the redo log buffer
    Redo Log Buffer is a piece of memory, it's not a file. No I/O.
    I don't believe Server process copies something directly from BUFFER CACHE to REDO LOG BUFFER.
    Ales
    The whole difference between a little boy and an adult man is the price of toys

  5. #15
    Join Date
    Dec 2000
    Location
    Ljubljana, Slovenia
    Posts
    4,439
    Originally posted by Sonia
    ....and server proscess is the one which writes from DBbuffer cache to the redo log buffer
    Sure, it also writes to buffer cache itself! But what has this to do with the subject discussed here? We were talking about physical writes.
    Jurij Modic
    ASCII a stupid question, get a stupid ANSI
    24 hours in a day .... 24 beer in a case .... coincidence?

  6. #16
    Join Date
    Apr 2001
    Location
    London
    Posts
    725

    Time to use OCP.

    This is what my official OCP training has taught me :

    CKPT (Writes to control file, Writes to data files)
    DBWn (Writes to data files)
    LGWR (Writes to log files)
    ARCn (Read log file, Writes archive log file, read/write control file)
    SVR (Reads data files)
    Once you have eliminated all of the impossible,
    whatever remains however improbable,
    must be true.

  7. #17
    Join Date
    Dec 2000
    Location
    Ljubljana, Slovenia
    Posts
    4,439

    Re: Time to use OCP.

    Originally posted by Sureshy
    This is what my official OCP training has taught me :

    CKPT (Writes to control file, Writes to data files)
    DBWn (Writes to data files)
    LGWR (Writes to log files)
    ARCn (Read log file, Writes archive log file, read/write control file)
    SVR (Reads data files)
    Well, it hasn't taught you everything, and what id has taught you is certanly not 100% complete. For example in the above quote it is not mentioned that LGWR is one of those processes that write to controlfiles.

    And most importantly, your official OCP training does not provide any answer to the original question of this thread: Which process is responsible for writing changed database names into the controlfile?
    Jurij Modic
    ASCII a stupid question, get a stupid ANSI
    24 hours in a day .... 24 beer in a case .... coincidence?

  8. #18
    Join Date
    Aug 2002
    Location
    Bangalore, India
    Posts
    405
    Originally posted by jmodic
    Yes, I would say so.

    I don't remember coming across the term "foreground process" in this context.
    It's user process not foreground process. Sometimes we invoke userprocess at the server end when we open sqlsessions at the server. This is the scenarion when we learn Oracle at home installing the Oracle s/w in our home PC-server
    -nagarjuna

  9. #19
    Join Date
    Aug 2002
    Location
    Bangalore, India
    Posts
    405
    Originally posted by jmodic
    Where did you hear that????? Server proces *can never* write to the database file. Never! It can only read from database file. It would be a shocking news to me if you could prove me wrong on that.

    Why would be any different with ALTER TABLESPACE COALESCE command? How does it differ with any other CREATE, ALTER, etc commands? No matter what command is executed, server process can only *read* from datafiles. All modifications that must be performed by server processare are executed in the buffer cache, and it is DBWR that performs any datafile writes associated with that.

    Realy, can you provide any explanation/URL regarding what is so special about ALTER TABLESPACE COALESCE in this context?
    I may be wrong. But, coelescing is done by a server process when we use this "alter tablespace ...." command.
    Last edited by nagarjuna; 10-25-2002 at 06:26 AM.
    -nagarjuna

  10. #20
    Join Date
    Dec 2000
    Location
    Ljubljana, Slovenia
    Posts
    4,439
    Originally posted by nagarjuna
    I may be wrong. But, coelescing is done by a server process when we use this "alter tablespace ...." command.
    Hey, lets make one thing clear: (Almost) everything in the database "is done" by server process - create table, drop table, craete tablespace, create database, alter database, alter system, .... you name it ... - everything is "done" by server process. But what does expression "done" mean here? It means that server process performs all necessery changes *in the buffer cache* and in the buffer cache alone! But those changed buffers will *never* be written to the datafiles by server process. Never! There are background processes that are responsible for that and only those processes will ever be able to do that. It's as simple as that, and I can't see why you've exposed ALTER TABLESPACE COALESCE as something special - its just one of those DDL commands, nothing special about it.

    Once again: the isue in this thread was: which process performs PHYSICAL WRITES into database (control) files.
    Jurij Modic
    ASCII a stupid question, get a stupid ANSI
    24 hours in a day .... 24 beer in a case .... coincidence?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  



Click Here to Expand Forum to Full Width