Search Postgresql Archives

Re: Vacuuming strategy

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Apr 30, 2014 at 10:49 AM, Elanchezhiyan Elango <elanelango@xxxxxxxxx> wrote:

Each table has 4 indices. The updates are to the indexed columns.
Here is the schema of a table: http://pastebin.com/ffu0dUjS All tables have this same schema except that some tables don't have a port column and so will have one less index
What indexes exist?  Are the updates to indexed columns?  

Sorry I was wrong when I said the updates are to the indexed columns. The updates are to the 'data[]' column which doesn't have any index.


In this case your updates might qualify for Heap-Only-Tuple (HOT) updates.  Those are nice because they don't bloat the indexes and because the space within the tables can be re-used faster, not needing an intervening vacuum in order to reuse it.  But to get that benefit, you do have to have enough space in the page to record the new tuple, and so it would benefit from a lower fillfactor.  But once that part of the table goes cold, then the lower fillfactor no longer is a benefit and justs wastes space.  That is where partitioning could come in handy.  Once the tuples have become old enough that they can't be updated any more by the app, just move them from the live partition to the cold partition and they get repacked in the process

But so far there is no indication you need to resort to such things.

Cheers,

Jeff

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Postgresql Jobs]     [Postgresql Admin]     [Postgresql Performance]     [Linux Clusters]     [PHP Home]     [PHP on Windows]     [Kernel Newbies]     [PHP Classes]     [PHP Books]     [PHP Databases]     [Postgresql & PHP]     [Yosemite]
  Powered by Linux