Subject: | |
From: | |
Reply To: | Lindholm, Scott - Broomfield, CO |
Date: | Thu, 5 Aug 1999 09:24:08 -0600 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
Stan,
I read the paper that you wrote on using SHOT and Debug. In that paper you
reference another paper, which you were going to write, that was going to
discuss what the different system processes do. I looked for that paper but
it was not with the other papers. I would appreciate it if you could help
me find it. We have quite a few system processes which are always running
and consuming a lot of CPU. I have not noticed this in other shops that I
have been in and I am trying to figure out if it is normal. I have a
listing from glance which shows the processes that I am talking about.
P2 SYS MANAGER.SYS 2 LOAD B142 4.3% 0.4 0 0.0
MISC
P3 SYS MANAGER.SYS 3 B100 4.5% 0.0 0 0.0
MISC
P5 SYS MANAGER.SYS 5 C152 1.1% 5.4 0 0.0
MSG
P41 SYS MANAGER.SYS 41 A 13 3.1% 0.0 0 0.0
P45 SYS MANAGER.SYS 45 A 13 2.3% 14.6 0 0.0
MSG
P46 SYS MANAGER.SYS 46 A 13 2.3% 13.8 0 0.0
MSG
P47 SYS MANAGER.SYS 47 A 13 2.4% 14.7 0 0.0
MSG
P62 SYS MANAGER.SYS 62 A 13 2.5% 17.0 0 0.0
MSG
P65 SYS MANAGER.SYS 65 A 13 1.5% 10.0 0 0.0
MSG
P72 SYS MANAGER.SYS 72 A 13 2.6% 17.8 0 0.0
MSG
P73 SYS MANAGER.SYS 73 A 13 2.6% 17.7 0 0.0
MSG
P74 SYS MANAGER.SYS 74 A 13 2.5% 17.5 0 0.0
MSG
P78 SYS MANAGER.SYS 78 A 13 3.2% 22.3 0 0.0
MSG
P79 SYS MANAGER.SYS 79 A 13 3.0% 20.9 0 0.0
MSG
Thanks,
Scott Lindholm
Sr Systems Administrator
Corporate Express
303-664-3264
[log in to unmask]
> -----Original Message-----
> From: Stan Sieler [SMTP:[log in to unmask]]
> Sent: Wednesday, August 04, 1999 4:53 PM
> To: [log in to unmask]
> Subject: Re: Pin 2 and Pin 11
>
> Hi all,
>
> Not to minimize what Doug wrote: :)
>
> ...
> > Not to minimize the value of SHOT, and for those not fortunate enough to
> > have such tools at their disposal, a quick stack trace of PIN 11 gives a
> > pretty good indication of what it is:
> >
> > :debug
> > DEBUG/iX C.05.08
> >
> > HPDEBUG Intrinsic at: a.00a2f698 hxdebug+$e4
> > $1 ($64) nmdebug > pin #11
> > $2 ($b) nmdebug > tr,i,d
> ...
>
> There's more info on the above in my 1993 paper:
> http://www.allegro.com/papers/whorun.html
>
> It's important to combine the "PIN" and "TR" command into one line:
>
> nmdebug> PIN #11; tr,i,d
>
> Why? Because "PIN" takes a snapshot of the registers of the process.
> The process may be executing ... if so, that snapshot gets older and
> older ... to the point where the "TR" command may get grossly incorrect
> data. Combining the two onto one line minimizes the timing window (but
> it doesn't eliminate it).
>
> (P.s.: don't forget that "#" ... it means "decimal". Otherwise,
> you're likely to be getting a trace of process #9 or process #17
> instead of #11.)
>
> BTW, I wrote the paper after writing an even earlier article for
> the HP Chronicle because I was tired of only having SHOT know
> how to do this... I wanted to encourage GLANCE, SOS, etc., to add
> the capability to get better process "name" information.
>
> --
> Stan Sieler [log in to unmask]
> http://www.allegro.com/sieler/
|
|
|