Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 5 Nov 2000 14:09:58 +0100 (CET)
From:      Maarten van Schie <AnEra@dds.nl>
To:        David Malone <dwmalone@maths.tcd.ie>
Cc:        Kent Stewart <kstewart@urx.com>, David Kelly <dkelly@hiwaay.net>, stable@FreeBSD.ORG
Subject:   Re: Strange latency? Was: 4.1.1-Stable
Message-ID:  <Pine.BSF.4.21.0011051324300.75325-100000@oT.o8.com>
In-Reply-To: <20001105110102.A74901@walton.maths.tcd.ie>

next in thread | previous in thread | raw e-mail | index | archive | help



On Sun, 5 Nov 2000, David Malone wrote:

> On Sun, Nov 05, 2000 at 03:34:41AM +0100, Maarten van Schie wrote:
> 
> If it doesn't seem to be DNS you could try one of the following to
> find out what is going on:
> 
> 1) While running the program use "top" in another window and see how
> 	much CPU time is being used and what it says in the "STATE"
> 	field for the processes that seem to take ages to start.
> 


After I start it it just sits down and does the following:

75337 root       2   0  4220K  2448K kqread   0:00  0.00%  0.00% pine

this lasts for about 2.5 minutes, then it comes into play and receives
about 0.05%CPU for some seconds after which it again goes 'kqread'
for about a minute. Then again a short peak in CPU % to finish with:

75337 root       2   0  4280K  2556K select   0:00  0.00%  0.00% pine

the usual...(random users experience the same thing.)


> 2) Run the program using ktrace by saying "ktrace program". Quit out
> 	of the program once it's done its going slow bit and use
> 	"kdump -R" or "kdump -T" to look for long delays.



I did it before but as 'ktrace -f ~/pine.ktr pine'.
Oeh..kdump works with other timestamps than I would:

973269692.747733 <- time when ktrace started
0.000444	 <- time since previous entry

I used root for this one, everything local(I only use it for cron
output).
So I don't know what kind of amounts to look for, but I came up with:

   978 pine     5.002428 RET   kevent 0
   978 pine     5.009282 RET   kevent 0
   978 pine     10.019485 RET   kevent 0
   978 pine     13.983712 RET   kevent 1
   978 pine     5.004701 RET   kevent 0
   978 pine     5.009152 RET   kevent 0
   
these are before Pine reads it's/the users configuration files and some
more peaks like these are shown. Later I see about 350 times:

   978 pine     0.000070 CALL  setitimer(0,0xbfbff0d8,0xbfbff0c8)
   978 pine     0.000047 RET   setitimer 0

after pine read the config files.

Later on:

978 pine     0.000231 CALL  write(0x1,0x82ae000,0x49)
   978 pine     0.000141 GIO   fd 1 wrote 73 bytes
       "\^[[48;1H\^[[K\^[[7m\^[[48;26H[Folder "INBOX" opened with 0
messages]\^[[27m\^[[\
        48;1H"
   978 pine     0.000067 RET   write 73/0x49
   978 pine     0.000061 CALL  setitimer(0,0xbfbff5c0,0xbfbff5b0)
   978 pine     0.000042 RET   setitimer 0
   978 pine     0.000074 CALL  sigprocmask(0x1,0,0x82a25bc)
   978 pine     0.000042 RET   sigprocmask 0
   978 pine     0.000092 CALL  select(0x1,0xbfbff5b8,0,0xbfbff538,0xbfbff638)
-> 978 pine     150.002556 RET   select 0
   978 pine     0.000280 CALL  getppid
   978 pine     0.000043 RET   getppid 211/0xd3
   978 pine     0.000097 CALL  gettimeofday(0xbfbff688,0)
   978 pine     0.000038 RET   gettimeofday 0
   978 pine     0.000055 CALL  gettimeofday(0xbfbff688,0)
   978 pine     0.000033 RET   gettimeofday 0
   978 pine     0.000041 CALL  gettimeofday(0xbfbff628,0)
   978 pine     0.000030 RET   gettimeofday 0

This is right before pine start to seek my mailfolder and some entries
later pine finishes.


> 
> David.
> 



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.21.0011051324300.75325-100000>