Reduce database size

  • Have been asked by a customer to reduce the database size to latest 10% of its data. Not sure how to achieve this. Any help regarding this would be very useful. Thanks in advance.

  • Your choices are pretty limited. You can remove data and then shrink the database. You can use storage compression if you're using Enterprise in SQL Server. You can look at a third party product like Red Gate SQL Storage Compress (disclosure, I work for Red Gate).

    In general, simply saying, make the database smaller, is somewhat difficult to answer. Why do they need this. What are you trying to achieve? Is it just a storage issue or something else going on?

    "The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
    - Theodore Roosevelt

    Author of:
    SQL Server Execution Plans
    SQL Server Query Performance Tuning

  • Its basically a storage issue and they are not in a position to expand the LUN being a UAT box

  • A further possibility could be hidden within the indexes. Do indexes have much fragmentation or are they reorganized regularly? Fragmented indexes can have a lot of impact on the their size .I have seen tables without doing maintenance for some weeks, after rebuilding they needed about 50 GB (!) less storage.

    Another step could be to analyze index usage. If indexes are not used they should be dropped as they need storage space too. (Please don't start dropping indexes now, analyzing is a little bit of work 😀 )

    Another possibility is thinking of the storage type. If your storage is for hig availability it might be expensive. Is there data that can be moved to an archive? Could the archive be placed on an less expensive storage as the archive might be "less secure"?

    Also check the size of the database log files and analyze if the size is really needed. Check if you can reduce log growing using a more frequently log backup.

    Check the size of the database files too. A database file might have lots of storage reserved but only less of it actually used (maybe the file growed for example due to index rebuild operations). Try to shrink the file to an appropriate size.

  • Grant is right about the indexes and transaction logs. Be sure the database is in simple mode, and run the DBCC ShrinkDatabase(<databasename>, 10) to recover any disk space that may be available from the log commitments. You mentioned that it is UAT. What are the testing requirements? Do you need to do a load test? Can you cull down the data, based on a date field? I just went throught this excercise on a development system. I asked our business users for the primary filter criteria, which turned out to be Member State. I applied a filter that focused on just 3 states for claims data. Once I deleted all the unwanted data and did a ShrinkDatabase, I reduced the database size on disk from 222GB to 74GB. :w00t:

  • I would NOT recommend use SHRINKDB "just for fun". Of course is clears all unused disk space but it also can have it's negative impact, from maximum fragmented indexes to performance, as for some operations the file has to grow again.

    Also setting recovery model to simple should be diskussed. It depends on how many data you lose if your system crashes. The full recovery model is no problem if you backup and thus empty the log file regularly.

  • ranganathleo (11/29/2012)


    Have been asked by a customer to reduce the database size to latest 10% of its data. Not sure how to achieve this. Any help regarding this would be very useful. Thanks in advance.

    Perhaps the easiest way to approach this disk space issue in UAT would be to periodically drop and restore the datbase from a backup.

    "Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho

  • to get it to 10% of it's original size, that means deleting 90% of the data? i don't think that's what you are really asking.

    besides the points above, (and compresison would be my first choice), find every table that is a HEAP, and put a clustered index on it; HEAP tables never release the space taken by deleted rows.

    Lowell


    --help us help you! If you post a question, make sure you include a CREATE TABLE... statement and INSERT INTO... statement into that table to give the volunteers here representative data. with your description of the problem, we can provide a tested, verifiable solution to your question! asking the question the right way gets you a tested answer the fastest way possible!

  • Sorry, I'm a newbie. What's a 'LUN', and what's a 'UAT'?

    Jim

  • WolfgangE (12/3/2012)


    I would NOT recommend use SHRINKDB "just for fun". Of course is clears all unused disk space but it also can have it's negative impact, from maximum fragmented indexes to performance, as for some operations the file has to grow again.

    Also setting recovery model to simple should be diskussed. It depends on how many data you lose if your system crashes. The full recovery model is no problem if you backup and thus empty the log file regularly.

    It is, however, a UAT box. Unless you're doing Point-in-time backups on your UAT box, there's no need to use any recovery mode other than SIMPLE. Also, a UAT box probably won't suffer the same amount of activity so it's not likely the log fie needs to be as large as the production box.

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.

    Change is inevitable... Change for the better is not.


    Helpful Links:
    How to post code problems
    How to Post Performance Problems
    Create a Tally Function (fnTally)

  • JimS-Indy (12/4/2012)


    Sorry, I'm a newbie. What's a 'LUN', and what's a 'UAT'?

    LUN is a Logical Unit Number on a SAN (a type of disk system separate from the server).

    UAT is "User Acceptance Testing".

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.

    Change is inevitable... Change for the better is not.


    Helpful Links:
    How to post code problems
    How to Post Performance Problems
    Create a Tally Function (fnTally)

  • A lot of people have very large "log" tables that mean nothing in UAT. Find them and truncate them before you do the recommended index and compression things the others have recommended.

    If that does do it, then you're going to have to pick and choose which data you want to keep. Such "Gold Sets" are a huge PITA in many ways.

    --Jeff Moden


    RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
    First step towards the paradigm shift of writing Set Based code:
    ________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.

    Change is inevitable... Change for the better is not.


    Helpful Links:
    How to post code problems
    How to Post Performance Problems
    Create a Tally Function (fnTally)

  • Lowell (12/4/2012)


    HEAP tables never release the space taken by deleted rows.

    this is new to me ..can you give me any article reference here ?

    -------Bhuvnesh----------
    I work only to learn Sql Server...though my company pays me for getting their stuff done;-)

  • Also UAT doesnt need archived or purged data (some historical table too) .. we can obsolete those tables too (+ their indexes too )

    -------Bhuvnesh----------
    I work only to learn Sql Server...though my company pays me for getting their stuff done;-)

  • Bhuvnesh (12/5/2012)


    Lowell (12/4/2012)


    HEAP tables never release the space taken by deleted rows.

    this is new to me ..can you give me any article reference here ?

    sure!

    take a look at this recent thread, where someones HEAP was holding 15 gig, but the real data was only 600Meg

    http://www.sqlservercentral.com/Forums/Topic1390660-149-1.aspx

    and the same citation i posted in that link for ease of clicking:

    http://sqlserverpedia.com/wiki/Heaps#Deletes_and_Heaps

    Deletes and Heaps

    When data is deleted from a heap using a DELETE statement, SQL Server will not release the space; it remains allocated to the heap. This leads to space bloat that wastes valuable resources. To address this problem, you can do any of the following:

    Lowell


    --help us help you! If you post a question, make sure you include a CREATE TABLE... statement and INSERT INTO... statement into that table to give the volunteers here representative data. with your description of the problem, we can provide a tested, verifiable solution to your question! asking the question the right way gets you a tested answer the fastest way possible!

Viewing 15 posts - 1 through 15 (of 22 total)

You must be logged in to reply to this topic. Login to reply