RE: db_file_multiblock_read_count causing full scans to takelonger?

  • From: "Kevin Closson" <kevinc@xxxxxxxxxxxxx>
  • To: "Kevin Lidh" <kevin.lidh@xxxxxxxxx>
  • Date: Tue, 19 Dec 2006 14:18:04 -0800

If this is UFS, you should be able to tunefs to set 
fs_maxcontig to limit the size of the gobbling an FS
read-ahead reads in. The wont control just how far a 
read-ahead reaches though. 

What you need (and I don't know how to track this on
HPUX) is visibility into the read-ahead blocks visited. All
Fses I've worked on track that. It is important to know
whether or not RA is getting triggered for stuff that 
never gets visited.

>>>-----Original Message-----
>>>From: Kevin Lidh [mailto:kevin.lidh@xxxxxxxxx] 
>>>Sent: Tuesday, December 19, 2006 1:22 PM
>>>To: kevinc@xxxxxxxxxxxxx
>>>Cc: oracle-l@xxxxxxxxxxxxx
>>>Subject: RE: db_file_multiblock_read_count causing full 
>>>scans to takelonger?
>>>In my case, it's filesystem.  Is there a parameter that 
>>>determines read-aheads?  I'll look for HP-UX but if you know 

Other related posts: