Summary - ps hang during boot

From: Richie Murphy III DSR ,Inc. (murphy@dsr.com)
Date: Wed Jun 15 1994 - 20:19:45 CDT


I'll restate the original problem.

   We have a problem which hopefully someone could help us with. We are
running Solaris 2.3 on a Sun690. The problem we are encoutering is the ps
command on boot and shutdown down "hangs" and eats up the cpu resources. The
console will not go past "syslog service started". In order to get it up we
can remotely log in and kill the ps process. The system then comes up
normally. The same senario occurs when you shut the system down. But we
cannot kill the ps because the network deamons are not long running. It has
to be manually powered off. This problem occured about 6 hours after an
install of additions disk drives. This same situtation also happened to a
Sparc1.

I would like to thank all the following people for their responses.
glenn
rmsun!rmorse@uunet.uu.net (Robert Morse)
Stuart.Little%ssu.stc.co.uk@comsys.Dialix.oz.au
saouli@math.ethz.ch
Casper Dik <casper@fwi.uva.nl>
arden@cse.unsw.edu.au
Timothy G. Smith <tgsmith@Sun.Com>

Below are excerpts from different emails stating the solutions.

------------------------------------------------------------------------------
The solaris FAQ says:

5.7) My machine hangs during the boot process. It seems related to ps.

        When the system boots, the first invocation of ps will try to
        recreate /tmp/ps_data. To this end ps scans the /dev tree.
        Under some circumstances, a loop exists in /dev and ps will
        run forever. Most of the time this loop is caused by the symbolic
        link /dev/bd.off. While this link usually points to /dev/term/b,
        it sometimes get truncated and points to /dev instead.

        Fix: rm -f /dev/bd.off; ln -s /dev/term/b /dev/bd.off

        Use truss(1) to determine whether this is real the cause of your
        problem.

    --- end of excerpt from the FAQ

The most recently posted version of the FAQ is available from
ftp.fwi.uva.nl in directory /pub/solaris

------------------------------------------------------------------------------
Why not edit the startup file and change the line that starts ps to
"truss -o /var/tmp/ps_log ps ...". Then have a look at the file to see
why ps is not running properly.

------------------------------------------------------------------------------
        Yes, I have seen this problem too. It seems like during the boor -r
process a symbolic link in the /dev directory get messed up. It usually is
the one that deals with the button & dial interface, /dev/bd0????. It should
point the ttya (Does not really matter where it points to but it must point
somewhere.)



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:09:03 CDT