Why does 'ps' hang system ?

Why does 'ps' hang system ?

Post by Tim Deev » Fri, 06 Mar 1992 11:57:08



I have a problem with a system* on bootup.  The system
is running v 3.1.5.  The problem started after the owner of the
machine attached an external disk, without shutting the system down.
Before I had a chance to look at it, he had removed the disk
from the machine (thinking that would fix his problem).
The machine would get as far as "Multi-user initialization
completed", and then just hang.  The "Multi-user..." is the
very last thing that gets done in /etc/rc, and the next thing
in inittab is rc.tcpip.  I commented out rc.tcpip and rc.nfs
in inittab, and now the machine comes up all the way.  I looked
at rc.tcpip to see what might be causing the problem, and the
first thing done there is to check if srcmstr is running by doing
a 'ps'.  Sure enough, if I do a 'ps' now it hangs the system.
Any ideas what would cause that to happen ?  One other thing:
I did a 'ls -l /dev' and that hangs the machine also (though
'ls /dev' does not).

Thanks for any help.

Tim Deeves

 
 
 

Why does 'ps' hang system ?

Post by Marc W » Sun, 08 Mar 1992 00:41:14




> I have a problem with a system* on bootup.  The system
> is running v 3.1.5.  The problem started after the owner of the
> machine attached an external disk, without shutting the system down.
> Before I had a chance to look at it, he had removed the disk
> from the machine (thinking that would fix his problem).
> The machine would get as far as "Multi-user initialization
> completed", and then just hang.  The "Multi-user..." is the
> very last thing that gets done in /etc/rc, and the next thing
> in inittab is rc.tcpip.  I commented out rc.tcpip and rc.nfs
> in inittab, and now the machine comes up all the way.  I looked
> at rc.tcpip to see what might be causing the problem, and the
> first thing done there is to check if srcmstr is running by doing
> a 'ps'.  Sure enough, if I do a 'ps' now it hangs the system.
> Any ideas what would cause that to happen ?  One other thing:
> I did a 'ls -l /dev' and that hangs the machine also (though
> 'ls /dev' does not).

Sounds like NIS (aka YP) isn't running or it's hosed.

Have you or anyone else made any changes in the order in which
stuff is started in the /etc/rc.* files?

When did the problem start?

Make sure that NIS is up and that your system can communicate with a
NIS server.  Start by making sure your network interfaces are up.  Then
verify that the portmapper is running on your system by rpcinfo -p.
NExt do a ypwhich to tell which server you're bound to.  Then do a
rpcinfo -u <server_name> ypserv.  Also do a rpcinfo -u <your_host>
ypbind.

Marc Wiz                                MaBell (512)838-4028

Yes that really is my last name.
The views expressed are my own.



 
 
 

Why does 'ps' hang system ?

Post by Dinah Mc Nu » Tue, 10 Mar 1992 02:22:43



writes:



> > I have a problem with a system* on bootup.  The system
> > is running v 3.1.5.  The problem started after the owner of the
> > machine attached an external disk, without shutting the system down.
> > Before I had a chance to look at it, he had removed the disk
> > from the machine (thinking that would fix his problem).
> > The machine would get as far as "Multi-user initialization
> > completed", and then just hang.  The "Multi-user..." is the
> > very last thing that gets done in /etc/rc, and the next thing
> > in inittab is rc.tcpip.  I commented out rc.tcpip and rc.nfs
> > in inittab, and now the machine comes up all the way.  I looked
> > at rc.tcpip to see what might be causing the problem, and the
> > first thing done there is to check if srcmstr is running by doing
> > a 'ps'.  Sure enough, if I do a 'ps' now it hangs the system.
> > Any ideas what would cause that to happen ?  One other thing:
> > I did a 'ls -l /dev' and that hangs the machine also (though
> > 'ls /dev' does not).

> Sounds like NIS (aka YP) isn't running or it's hosed.

> Have you or anyone else made any changes in the order in which
> stuff is started in the /etc/rc.* files?

> Make sure that NIS is up and that your system can communicate with a
> NIS server.  Start by making sure your network interfaces are up.  Then
> verify that the portmapper is running on your system by rpcinfo -p.
> NExt do a ypwhich to tell which server you're bound to.  Then do a
> rpcinfo -u <server_name> ypserv.  Also do a rpcinfo -u <your_host>
> ypbind.

Another thing you can do is boot the system single user mode and execute
the scripts by hand. I have found this useful on other UNIX systems
so you can watch the errors was the script executes. I typo in one rc.*
file can cause unpredictable results (usually in the rest of the file
being ignored.)
--

Dinah McNutt                                    Pencom Software, Inc.

 
 
 

1. SIGFPE with 'top' or 'ps -u' -- why ???

Hi there,

I'm working woth Linux 1.3.x (1.3.88 at this time).

Everytime I call 'ps' with the user option (-u) or 'top',
which lists user dependent, too, the process dies getting a
FPE signal (floating point exception).

Has someone got a solution for this problem ?
I tried to trace the process hopefully to identify the problem,
but I didn't.

thanx Alexander Strauss

2. sunrpc port 111

3. why would 'cp' or 'ls -al' command hangs on Solaris 2.7 ?

4. Problem making alternate users.

5. Is e2label 'dangerous' when done on a 'live' file system?

6. OpenWIN performance over Ethernet

7. hung after 'exit()' is called...what's it doing?

8. thanks & help: BIND

9. The system hangs when it's about to start 'init'

10. Limiting 'w', 'top' and 'ps'

11. How to hide from 'who', 'finger', 'ps' commands?

12. 'ps' command hangs

13. 'ps' command hangs in FreeBSD 3.3R?