At the very end is the original questions I asked.
Summarizing:
Problem 1: I received various solutions on how to try and debug where it
was hanging.
1. boot into level 1 and run each script by hand until identifying the
troublesome one. Could add an echo "starting--..." in each /etc/init.d
script so it identifies itself. I thought this would take a lot of time but
since I was pretty sure it had something to do with network issues I only
added the echo to certain scripts.
2. possible hardware, try fscking disks or running test all from the PROM.
I did do the test all and everything checked out ok.
3. the hang is due to no network (this is getting closer), comment out any
possible troublesome items from inetd
4. David Evans indicated that it might be a problem with NIS having time to
bind a known bug. BINGO....it turned out not to quite be this particular
error but it was NIS related.
I figured it out by using the echo starting statement and finding that it
was stopping at the beginning of inetsvc. I had to put some additional
comments in but eventually narrowed it down to the fact that it wasn't
finding an NIS server. I commented out the startup of NIS in rpc and bingo
the machine booted. Then if I started ypbind by hand in broadcast mode all
was well.
I had forgotten to mention that I had tried to configure this machine as a
backup NIS server to my 2.6 box running as the NIS master. It never seemed
to be quite happy but I wasn't overly worried because I have plans to
upgrade this box to 2.6 also soon. I think the problem had to do with this
machine trying to bind to itself and not being able to. The only way I
could get it to boot with NIS running was to turn it back into an NIS
client. It turns out this is a known bug on SunSolve (bug ID 4162092).
Problem 2 - this machine couldn't boot simply because the other machine was
down and it was hung trying to NFS mount.
Problem 3 - turned out to be twofold
1. following some Sun docs I was trying to use /dev/aliases/floppy0. That's
not valid you have to use /vol/dev/diskette0 instead.
2. Also, for some reason volmgt would not mount the floppy. I had to stop
volmgt and mount the floppy by hand to get access to it.
Thanks to the following people who helped.David Evans, Sweth Chandramouli,
Sebastian "Benno" Benoit, Imre Kolos, Kevin Inscoe, Bill Armand, Jeff
Wimmer, Ganeshan and Dennis Martens
Thanks,
Lisa
----- Original Message -----
From: Lisa Weihl <lweihl@cs.bgsu.edu>
To: Sun Managers mailing list <sun-managers@sunmanagers.ececs.uc.edu>
Sent: Monday, July 12, 1999 10:10 PM
Subject: 3 machine problems (2 won't boot/1 floppy)
: When it rains it pours. I have 3 machines with various problems. The good
: news is for the two machines that won't boot I have replacement machines
: that are running. Ha, now those users will be forced to try the new
: equipment!!
:
: I hope you can all help me with these questions though...
:
:
: 1. SparcServer 5 running Solaris 2.5.1 serving as a telnet login machine
: for users. We've been having problems with it just crashing. You just
walk
: up to it and the screen is froze or blank. The only way to reboot is the
: power switch. I've never been able to track any logs from what has
: happened. Should I be able to find something? Next, now on reboot it
hangs
: right after starting rpc services. It lists all the rpc services and then
: says done and just sits there. I know the next thing in the boot sequence
: is setting the netmask so what exactly does this mean? How do I go about
: debugging what is failing in the boot sequence? This machine does mount
: filesystems off a machine that is up and running and it shares a
filesystem
: to the other machine that won't boot(see #2).
:
: 2. The second machine, a SparcStation 2 running Apache and running
Solaris
: 2.5.1, won't boot. Hangs right after it says starting inetd. It does
mount
: a filesystem off the machine in #1 but it seems like it should get further
: and directly complain about the mount from what I remember if that was the
: problem. The other thing to be noted on this machine is that it has been
: slow to answer telnet requests over the last few months. It was almost
: like it was in a sleep mode. It might take a minute to answer the first
: request but then it would answer the next one right away. I'm no
: networking guru and wasn't sure how to debug this. A netstat -a never
: showed an inappropriate amount of connections or connections in a
time_wait
: state. Tonight it wouldn't answer the telnet request at all and when I
: rebooted it hung as mentioned.
:
: 3. The last problem is with an Ultra 5 running Solaris 2.6. It doesn't
: seem to want to be able to mount a floppy. I can format a floppy but
cannot
: get a filesystem on it. A vold troubleshooting doc from sunsolve says to
: try newfs /vol/dev/aliases/floppy0 to put a fs on the floppy. I have a
: /vol directory but nothing underneath it so it doesn't like that command.
: This problem arose because I'm trying to install IBM's APL2 for Sun
Solaris
: on the box and it comes on floppies which I don't seem to be able to read.
: I know there are all kinds of problems with vold and mounting floppies.
My
: cd management works fine. I couldnt' find anything specific to this for
: Solaris 2.6 with a search of sunsolve. I did find many complaints about
: problems on Ultra 5's with floppies but not this specific one. Any ideas?
:
: TIA for help, will summarize,
: Lisa
:
******************************************************************
Lisa Weihl, System Administrator E-mail: lweihl@cs.bgsu.edu
Department of Computer Science Office: Hayes 225
Bowling Green State University Phone: (419) 372-0116
Bowling Green, Ohio 43403-0214 Fax: (419) 372-8061
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:13:24 CDT