DBAsupport.com Forums - Powered by vBulletin
Page 1 of 2 12 LastLast
Results 1 to 10 of 12

Thread: archiving problem, very urgent

  1. #1
    Join Date
    Apr 2002
    Posts
    291
    Hi gurus,
    I've a strange problem. My log_buffer size is set to 5M, and my log file size is 20M. When i checked the archive logs sizes, its varying from one file to another. some are around 10M, some are only 1KB. like different sizes. I didn't put any checkpoint interval in my parameter file. I need to have my logfile generated in the same size. can any one please help me. This is on production box. It's very urgent.

    Due to this, my tape backup is not running. Its again a different scenario.

    Thanks in advance
    PNRDBA

  2. #2
    Join Date
    Oct 2000
    Location
    Charlotte, USA
    Posts
    330
    Hello,
    what is the database version and OS?.
    if it is 9i can you change the log_buffer size????.
    Thanks.
    Thanigai.

  3. #3
    Join Date
    Apr 2002
    Posts
    291
    Hi Mr.Vasan
    This is Nagraj here, i mailed you on the other day. Its 9i on 7.2 Linux. my problem is solved. The problem was, my log buffer size was around 200 KB, then i changed it to 5 M. Now its working fine. Is there any formula to fix the log buffer with respect to logfile?
    PNRDBA

  4. #4
    Join Date
    Apr 2002
    Location
    Phoenix, AZ
    Posts
    175
    I dont get this, how can Log_Buffer Size matter to Archive Log Sizes ??

    It is possible that you have Redo Logs with different sizes.
    Sridhar R Patnam

  5. #5
    Join Date
    Dec 2000
    Location
    Ljubljana, Slovenia
    Posts
    4,439
    Of course LOG_BUFFER size can not influence the size of archived log files in any way! Only the size of online redo log files and forced log switches (like "ALTER SYSTEM SWITCH LOGFILE;") can influence the size of archived redo logs.

    And BTW, LOG_BUFFER of 5 MB is nothing but waste of RAM, as have been explained many times in this forum...
    Jurij Modic
    ASCII a stupid question, get a stupid ANSI
    24 hours in a day .... 24 beer in a case .... coincidence?

  6. #6
    Join Date
    Jan 2002
    Location
    Netherlands
    Posts
    1,587
    Originally posted by jmodic
    And BTW, LOG_BUFFER of 5 MB is nothing but waste of RAM, ..
    Unless your intention was to axe the MTTR.
    Tarry Singh
    I'm a JOLE(JavaOracleLinuxEnthusiast)
    TarryBlogging
    --- Everything was meant to be---

  7. #7
    Join Date
    Dec 2000
    Location
    Ljubljana, Slovenia
    Posts
    4,439
    Originally posted by Tarry
    Originally posted by jmodic
    And BTW, LOG_BUFFER of 5 MB is nothing but waste of RAM, ..
    Unless your intention was to axe the MTTR.
    The size of LOG_BUFFER has absolutely no relevance on the MTTR! I suppose you had redo log file size in mind, not LOG_BUFFER.
    Jurij Modic
    ASCII a stupid question, get a stupid ANSI
    24 hours in a day .... 24 beer in a case .... coincidence?

  8. #8
    Join Date
    Jan 2002
    Location
    Netherlands
    Posts
    1,587
    U mean to say that if your redo log buffer size is relatively small, it will have no effect on the recovery process and this prolonging the time to recover?
    Tarry Singh
    I'm a JOLE(JavaOracleLinuxEnthusiast)
    TarryBlogging
    --- Everything was meant to be---

  9. #9
    Join Date
    Jun 2001
    Location
    Helsinki. Finland
    Posts
    3,938
    Originally posted by Tarry
    U mean to say that if your redo log buffer size is relatively small, it will have no effect on the recovery process and this prolonging the time to recover?
    MTTR (approx.) for crash recovery can be computed using the following formula:

    MTTR = Initialization_Time + Number_of_Files * File_Open_Time + 2 * Number_of_Log_Blocks * Log_Read_Time_per_Block +
    Number_of_Data_Blocks_to_Recover * (Recovery_Read_Time_per_Data_Block + Write_Time_per_Data_Block)
    Oracle Certified Master
    Oracle Certified Professional 6i,8i,9i,10g,11g,12c
    email: ocp_9i@yahoo.com

  10. #10
    Join Date
    Jan 2002
    Location
    Netherlands
    Posts
    1,587
    Ah Thanx for the formula julian but my Q was if your log buffer size is say default(that is 4 times the block size) but you have frequently long running transactions and a lot of them.
    Your instance crashes.. you restart the db and it takes a hell of a longtime.
    your MTTR statistics give you a certain figure.
    Would it not expedite the recovery time if your log buffer size was adjusted?
    That's what I wanted to confirm?

    Cheers!!!

    Tarry
    Tarry Singh
    I'm a JOLE(JavaOracleLinuxEnthusiast)
    TarryBlogging
    --- Everything was meant to be---

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