[Nut-upsuser] Slaves under Squeeze

[Rob Donovan]

hi,
     I recently upgraded my Debian Lenny system to Squeeze, and have 
been happily running nut as a MODE=standalone system talking to my 
Cyberpower 1500 avr lcd via the usbhid-ups driver for a couple of weeks 
or so.  No problems.

    Now I'm trying to hang another machine off the same ups as a slave, 
and it isn't working.  The problem I'm getting is best illustrated by 
what happens when I run this on the server:

upsd -DDDDDDDDD
Network UPS Tools upsd 2.4.3
   0.000000     listen_add: added 127.0.0.1:3493
   0.000070     listen_add: added 192.168.1.104:3493
   0.000117     setuptcp: try to bind to 192.168.1.104 port 3493
   0.000204     setuptcp: bind: Cannot assign requested address
   0.000226     not listening on 192.168.1.104 port 3493

Or when I run this on the slave :

/etc/init.d/nut start
Starting Network UPS Tools: upsmon.

Broadcast Message from nut at mac2
        (somewhere) at 16:33 ...

Communications with UPS cp1500avr1 at 192.168.1.102 lost


Broadcast Message from nut at mac2
        (somewhere) at 16:33 ...

UPS cp1500avr1 at 192.168.1.102 is unavailable

I presume the problem on the slave is a consequence of what's happening 
on the server... though I'm not entirely sure. 
Can anyone shed any light on this?
Details of my setup, including /etc/hosts.allow/deny files, netstat 
results and a few things I've already tried, follow.

Thanks
Rob Donovan


I installed nut with apt-get install, and I'm running version 
2.4.3-1.1squeeze1

The server machine, mac1, has IP address 192.168.1.102
The slave machine, mac2, has IP address 192.168.1.104

On the server :

nut.conf:
MODE=netserver

upsd.conf:
LISTEN 127.0.0.1
LISTEN 192.168.1.104

upsd.users:
[mac1mon]
   password = pass1
   upsmon master

[mac2mon]
   password = pass2
   upsmon slave

Other server setup files just the same as they were under MODE=standalone.

On the slave:

nut.conf
MODE=netclient

upsmon.conf
MONITOR cp1500avr1 at 192.168.1.102 1 mac2mon pass2 slave

Other setup files not relevant (I think).

/etc/hosts.allow on both machines :

ALL: LOCAL
ALL: 127.0.0.1
ALL: 192.168.1.

/etc/hosts.deny on both machines :

ALL: ALL

netstat -a -n shows that port 3493 is not in use on either machine.

Both machines running 2.6.32-5-amd64 kernel.
Server was upgraded from Lenny to Squeeze.
Slave is a new Debian 6.0.2.1-amd64 install.
Both machines can ping each other, ssh into each other, and nfs is 
exporting directories from the server to the slave.
I don't run any firewalls (that I know of).

I've already tried commenting out the contents of /etc/hosts.deny : no 
effect.
If I remove the 192 line from upsd.conf on the server, it will start, 
but the slave results are the same.
I've tried specifying 3493 for the 127 address and 3494 for the 192 
address in case that was a conflict : same results.