Subject: Re: Extremely slow IO response times + panic on 3.0_BETA
To: Valtteri Vuorikoski <vuori@sci.fi>
From: Dan LaBell <dan4l-nospam@verizon.net>
List: port-i386
Date: 07/13/2005 19:45:27
On Jul 13, 2005, at 6:11 PM, Valtteri Vuorikoski wrote:

> mlelstv@serpens.de (Michael van Elst) writes:
>
>> vuori@sci.fi (Valtteri Vuorikoski) writes:
>>
>>> I've been running 3.0_BETA CVS'd Jun 30 on a P4 box with two PATA 
>>> drives
>>> mirrored with raidframe. When doing disk io that involves heavy 
>>> metadata
>>> operations, (ie. tar xzf pkgsrc.tar.gz) sometimes disk IO response 
>>> gets
>>> extremely slow. It may take several minutes to do ls -l /usr/bin and
>>> another 30 seconds to get the prompt back once the listing is 
>>> finished.
>>
>> Yes, that's normal. The disks are almost only writing and there is
>> no memory for the filecache. But this was already the case for 2.0.
>
> I don't remember the effect being quite as severe on 2.0; maybe I just
> haven't tried the quite the same things.
>
> While I can see the mechanism behind it, the near-complete-hanging of
> the system is not a very pleasant experience.
>
>  -valtteri
>
I recently experienced something like this, but w/o any metadata 
operations, just dd on the
raw device.
I just got a 149 GB drive , it's destined to replace a 80 gig in a fw 
external drive,
but I decided to test it out on my pc first.   It seemed completely 
zeroed (fresh) , but
to verify that ,and to do a read-test, I did  a: dd if=/dev/rwd1d 
bs=128k | sum
Afterwards, there would be about a 3 second window, where I could type, 
maybe switch vt's,
but, then a long hang, which seemed last around a minute, possibly 
longer, before any kind of responsiveness was returned.  I'm wondering 
if the affect could be tuned away?
( I'm running 2.0.2 )