Very Computer

Hi,

On a 2.0.30 system, I am using ipx_configure and ncpmount to access
files on a local Novell system.  Recently I started process accounting
running on my system, and I noticed that I was getting a lot of
entries in the accounting log like:

  Userid Group  TTY  Command          Utime Stime Runtime Exit Proc start time
    root  root   ?   kerneld           0.00  0.00    0:00  126 14:33:10
    root  root   ?   sh                0.01  0.00    0:00  126 14:33:10
    root  root   ?   kerneld           0.00  0.00    0:00  126 14:33:08
    root  root   ?   sh                0.01  0.00    0:00  126 14:33:08
    root  root   ?   kerneld           0.00  0.00    0:01  126 14:33:05
    root  root   ?   sh                0.02  0.00    0:00  126 14:33:06
    root  root   ?   kerneld           0.00  0.01    0:00  126 14:33:04
    root  root   ?   sh                0.01  0.00    0:00  126 14:33:04
    root  root   ?   kerneld           0.00  0.00    0:00  126 14:33:02
    root  root   ?   sh                0.01  0.01    0:00  126 14:33:02
    root  root   ?   kerneld           0.00  0.00    0:01  126 14:32:59
    root  root   ?   sh                0.02  0.00    0:00  126 14:33:00

I had been running a kernel configured with IPX as a module, so I
re-compiled the kernel to have IPX built in, but I still get the same
sequence of kerneld/sh processes exiting every two (or so) seconds.
However, if I ncpumount the Novell file system, these message stop.

I've browsed some documentation, but I can't find anything out about
this.  (Yeah, I probably missed the correct doc, but...)

Can anyone tell me why all these kerneld/sh processes are happening?
Any way to prevent it (other than unmounting the Novell files)?

Thanks.

                        Jim Diamond

Quote:

> Hi,

> On a 2.0.30 system, I am using ipx_configure and ncpmount to access
> files on a local Novell system.  Recently I started process accounting
> running on my system, and I noticed that I was getting a lot of
> entries in the accounting log like:

>   Userid Group  TTY  Command          Utime Stime Runtime Exit Proc start time
>     root  root   ?   kerneld           0.00  0.00    0:00  126 14:33:10
>     root  root   ?   sh                0.01  0.00    0:00  126 14:33:10
>     root  root   ?   kerneld           0.00  0.00    0:00  126 14:33:08
>     root  root   ?   sh                0.01  0.00    0:00  126 14:33:08
>     root  root   ?   kerneld           0.00  0.00    0:01  126 14:33:05
>     root  root   ?   sh                0.02  0.00    0:00  126 14:33:06
>     root  root   ?   kerneld           0.00  0.01    0:00  126 14:33:04
>     root  root   ?   sh                0.01  0.00    0:00  126 14:33:04
>     root  root   ?   kerneld           0.00  0.00    0:00  126 14:33:02
>     root  root   ?   sh                0.01  0.01    0:00  126 14:33:02
>     root  root   ?   kerneld           0.00  0.00    0:01  126 14:32:59
>     root  root   ?   sh                0.02  0.00    0:00  126 14:33:00

> I had been running a kernel configured with IPX as a module, so I
> re-compiled the kernel to have IPX built in, but I still get the same
> sequence of kerneld/sh processes exiting every two (or so) seconds.
> However, if I ncpumount the Novell file system, these message stop.

-jr
--
Johannes Reisinger                           Tel: +43/1/8100606-329
Frequentis Ges.m.b.H. Vienna, AUSTRIA        Fax: +43/1/8100606-399

Quote:>On a 2.0.30 system, I am using ipx_configure and ncpmount to access
>files on a local Novell system.  Recently I started process accounting

Quote:>I had been running a kernel configured with IPX as a module, so I
>re-compiled the kernel to have IPX built in, but I still get the same
>sequence of kerneld/sh processes exiting every two (or so) seconds.
>However, if I ncpumount the Novell file system, these message stop.

Try running kerneld with the debug option ('kerneld debug').

ncpfs appears to be built with a hardcoded path '/sbin/nwmsg' which
gets passed to kerneld upon receipt of a netware message.  Kerneld
launches (or tries to launch) /sbin/nwmsg.

I was noticing a lot of kerneld zombies until I created a link from
/sbin/nwmsg to /usr/sbin/nwmsg (where the ncputils default nwmsg).

sdb
--
Do NOT send me unsolicited commercial e-mail (UCE)!
Watch out for munged e-mail address.
User should be sylvan and host is *highway.net.

I had a similar problem when I included IPX support in my
kernel.  I was getting a stream of log msgs that
it couldn't find a daemon that it was constantly trying to run.
The problem was solved by installing SAMBA