Subject: Re: from cheetah to snail
To: Frank Kardel <Frank.Kardel@Acrys.COM>
From: Patrick Welche <firstname.lastname@example.org>
Date: 02/04/2004 18:00:18
On Wed, Feb 04, 2004 at 02:29:44PM +0000, Frank Kardel wrote:
> Hi *,
> i have had similar observations (see eMail to current-users with subject
> "ps crawling...").
> My system is also a larger main memory system (1GB). I observed some
> unsusual things last night comparable to Patrik's findings.
> Here is a list:
> - I found pagedaemon running wild for some time so the top
> showed 98% CPU usage.
> - vmstat bufs is high cpu and cannot keep up with the usual display interval
> - ps take about 100ms per process (it is faster shortly after re-boot)
> - ps has system times in the > 98% range (it takes about 15s system time).
> Could this be related to the other thread in -current lately about
> pmap 0 showing many memory segments/fragmentation ?
% uname -sr && uptime && pmap 0 | wc -l
3:27PM up 22:16, 6 users, load averages: 1.33, 1.62, 1.63
% Wed Feb 4 16:34:17 GMT 2004
As in it took 1:07 to do the pmap.. At least that gives 2 options:
- try compiling with NEW_BUFQ_STRATEGY
- try code from before 2004/01/29 12:06:02; reverting workaround for