DB2 Basics: What Filesystems does my Database Actively Use?

You may also like...

2 Responses

  1. Rob says:

    Sorry – appreciate I’m dragging up an old post here – I was wondering if you had an opinion on the best storage block size to configure for a filsystem for DB2. I am working on a storage performance issue with a client running Linux DB2 servers who has a 512KB block size configured. The default is 4KB I believe, so that seems high, but I can’t find the person who originally configured that to ask why it was chosen. The only other related information I have at the moment is that they told me DB2 commits writes from RAM cache in 4MB chunks. I’m curious if this is something you have any recommendations on?

    Thanks & great blog

    • Ember Crooks says:

      The storage block size generally needs to be a multiple of the Db2 page sizes being used. Potential page sizes are 4K, 8K, 16K, or 32K. While a page is the smallest unit of data that Db2 works with, often pages are grouped into extents, and extents are the smallest size of data Db2 can read from disk (while still working with individual pages in memory). 512K is kind of a default block size for OLTP database servers. For OLTP database servers, we often want to read very small amounts of data – the prioritization is singleton-row accesses. Larger blocks may make more sense for Data Warehousing or Analytics systems, which tend to read larger chunks of data into memory.

      So the size of “chunks” Db2 is using depends heavily on what page sizes of the most active table spaces are and the extent sizes for those table spaces. The storage block size should be a multiple of page size * extent size.

Leave a Reply

Your email address will not be published. Required fields are marked *