Re: kernel 2.6.13 buffer strangeness

[Anthony Wesley <awesley@acquerra.com.au>]

Nate Diller wrote:
> just found the culprit.  guess i should have read the code the first 
> time.  get_dirty_limits() in drivers/block/page_writeback.c has a 
> hard-coded upper limit to dirty_ratio.  it's capped to half of the 
> unmapped pages, so maybe 30-40% of your system's memory.  so if you are 
> brave, just remove the "/ 2" parts from the 'if (dirty_ratio > 
> unmapped_ratio / 2) dirty_ratio = unmapped_ratio / 2;' check, and you 
> can have all the OOM goodness you want.

Excellent. OOM here I come.

> i really recommend you focus on getting better disk bandwidth, you stand 
> to gain a lot more from that approach.  i presume you're on ext3; 
> perhaps you should try reiser4 or xfs, they are more likely to meet your 
> disk bandwidth requirements.

Yep, pursuing this as well, also looking to add more RAM to the machine. At this stage
I was just trying to understand the numbers that I was seeing, so I could work out the
best way to proceed.

I'm using ext2 at the moment, on the assumption that the journal would cost 
me a bit of performance so I left it out :-)

I'll certainly try the other filesystems as you suggest.

Thanks again, Anthony

-- 
Anthony Wesley
Director and IT/Network Consultant
Smart Networks Pty Ltd
Acquerra Pty Ltd

Anthony.Wesley@acquerra.com.au
Phone: (02) 62595404 or 0419409836