[792] in Zephyr_Bugs
Problem with Zephyr 2.03
daemon@ATHENA.MIT.EDU (gantt (g.) edmiston)
Wed Jul 3 12:33:27 1996
Date: Wed, 3 Jul 1996 12:32:00 -0400
From: "gantt (g.) edmiston" <gantte@nortel.ca>
To: bug-zephyr@MIT.EDU
Cc: "william (w.) pfister" <pfister@nortel.ca>, ghudson@MIT.EDU
I've just built and installed Zephyr 2.03 on my HP 9000/700 systems.
Specifically the systems are:
HP-UX b4pph12d A.09.05 A 9000/730 2015373094 47.194.32.78
HP-UX b4pph056 A.09.03 A 9000/750 2002470738 47.194.35.87
Both systems are on the same subnet of a Class A network.
Our netmask is 255.255.240.0.
As you can see, the two machines are running slightly different versions
of the HP-UX operating system. Regardless, I was able to configure and
make the Zephyr executables on both, with just minor problems.
PROBLEM: I cannot get the two zephyrd server processes to recognize
each other.
We built Zephyr WITHOUT kerberos or hesiod support. Therefore, I have
a /usr/local/etc/zephyr/server.list file on both servers. They both
have identical information, which is:
root@b4pph12d-> cat /usr/local/etc/zephyr/server.list
47.194.32.78
47.194.35.87
I can start both zephyrd processes.
I can start zhm processes on client HP systems running 9.05.
I start the host manager like this:
/usr/local/sbin/zhm b4pph12d b4pph056
I have also tried starting it like this:
/usr/local/sbin/zhm 47.194.32.78 47.194.35.87
By themselves, both servers running the zephyrd process work independently
and appear to work fine.
Regardless, each server thinks the other is dead. Note the output from
the zstat command:
root@b4ppha6c-> zstat
Hostmanager stats: b4ppha6c
HostManager protocol version = ZEPH0.2
Current server = b4pph12d
Items in queue: 0
Client packets received: 16
Server packets received: 17
Server changes: 1
Version: $Id: zhm.c,v 1.58 1995/07/07 22:17:25 ghudson Exp $
Looking for a new server: no
Time running: 0 days, 01:10:51
Size: -1
Machine type: hppa1.1-hp-hpux9.05
Server stats: b4pph12d
Server protocol version = ZEPH0.2
Current server version = Zephyr Server $Revision: 3.25 $: (Tue Jul 2 14:58:12 EDT 1996) the_zephyr_builder@b4pph12d/
Packets handled: 18 pkts
Uptime: 0 days, 01:15:52
Server states:
47.194.32.78/SERV_UP
47.194.35.87/SERV_DEAD
If I issue a "zctl new_server", then zstat shows the new server as being
alive and the original as being dead.
If I sent a SIGFPE signal to both zephyrd processes, they think the
other server is dead. Can anyone please help me figure out how to
make these two zephyrd processes speak to each other?
I can be reached:
Gantt Edmiston
919-905-4991
or
gantte@nortel.ca
thanks!
Gantt Edmiston
NORTEL
DMS10 Tools development
400 Perimeter Park
Morrisville, NC, USA
