db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mike Matrigali (JIRA)" <j...@apache.org>
Subject [jira] Updated: (DERBY-4050) Multithreaded clob update causes growth in table that does not get reclaimed
Date Fri, 06 Feb 2009 21:21:01 GMT

     [ https://issues.apache.org/jira/browse/DERBY-4050?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Mike Matrigali updated DERBY-4050:
----------------------------------


note - the column is NUMUNFILLEDPAGES  pages not NUMFILLED.   See below for comments from
the diag SpaceTable javadoc about what the fields mean.  

    <P>The SpaceTable virtual table has the following columns:
    <UL>
    <LI>CONGLOMERATENAME varchar(128) - nullable.  The name of the conglomerate,
    which is either the table name or the index name. (Unlike the
    SYSCONGLOMERATES column of the same name, table ID's do not appear
    here).</LI>
    <LI>ISINDEX SMALLINT - not nullable.  Is not zero if the conglomerate is an
    index, 0 otherwise.</LI>
    <LI>NUMALLOCATEDPAGES bigint - not nullable.  The number of pages actively
    linked into the table.  The total number of pages in the file is the
    sum of NUMALLOCATEDPAGES + NUMFREEPAGES.</LI>
    <LI>NUMFREEPAGES bigint - not nullable. The number of free pages that
    belong to the table.  When a new page is to be linked into the table the
    system will move a page from the NUMFREEPAGES list to the NUMALLOCATEDPAGES
    list.  The total number of pages in the file is the sum of
    NUMALLOCATEDPAGES + NUMFREEPAGES.</LI>
    <LI>NUMUNFILLEDPAGES bigint - not nullable.  The number of unfilled pages
    that belong to the table. Unfilled pages are allocated pages that are not
    completely full. Note that the number of unfilled pages is an estimate and
    is not exact. Running the same query twice can give different results on
    this column. </LI>
    <LI>PAGESIZE integer - not nullable.  The size of the page in bytes for
    that conglomerate.
    </LI>
    <LI>ESTIMSPACESAVING bigint - not nullable.  The estimated space which
    could possibly be saved by compressing the conglomerate, in bytes.</LI>
    </UL>

> Multithreaded clob update causes growth in table that does not get reclaimed
> ----------------------------------------------------------------------------
>
>                 Key: DERBY-4050
>                 URL: https://issues.apache.org/jira/browse/DERBY-4050
>             Project: Derby
>          Issue Type: Bug
>          Components: Store
>    Affects Versions: 10.2.2.0, 10.3.3.0, 10.4.2.0, 10.5.0.0
>            Reporter: Kathey Marsden
>         Attachments: ClobGrowth.java, derby.log.growth, derby.log.nogrowth
>
>
> Doing a multithreaded update of a Clob table causes table growth that does not get reclaimed
except by compressing the table.  The reproduction has a table with two threads. One  thread
 updates row 1 repeatedly with 33,000 character clob. The other thread updates row 2 with
a small clob, "hello".  The problem occurs back to 10.2 but seems much worse on trunk than
10.2.   The trunk database grew to 273MB on trunk after 10000 updates of each row. The 10.2
database grew only to 25MB.  If the update is synchronized there is no growth.
> I will attach the repro.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message