Citeren Maurice Volaski <[email protected]>:
Please don't mistake warning messages with (fatal) errors. Starting
with nut-2.4.0, these messages should only be displayed in debug
mode, so I'm surprised you're seeing them in nut-2.4.1.
It doesn't seem this way from my reading of the source code:
upslogx(LOG_ERR, "[%s] %s: Error in packet: %s",
This is a different message from what you reported before. These where
logged with 'nut_snmp_get' in them and these lines should now be gone
(unless running in debug mode 2 or higher).
From what I can tell, that's a regular log message of an error, not
a debug mode message, which would use either "upsdebugx" or "debug",
or a warning, which would use upslogx with "LOG_WARNING".
Indeed, but this is in the lines you posted.
Upon startup, the snmp-ups driver will query the UPS for all the
OID's the driver supports. The ones which are not supported by the
UPS, will
In addition, the errors are continually output to syslog; they don't
just appear once and stop.
It looks like this is a different problem than what you mentioned
before. Please be specific.
generally result in an error message from the NetSNMP library that is
used. There is nothing we can do about that and it is *not* an error.
Perhaps you could change LOG_ERR to LOG_WARNING and perhaps you can
ignore it after the first time it appears.
We already do rate limiting for repeating errors. After startup, these
really shouldn't appear, which means there is probably something wrong
(and we want to know about that).
> Anyway, now that the script is starting, I'm seeing "failed - got
> [ERR ACCESS-DENIED]" errors from upsmon, and I don't know why.
See 'man 8 upsd', 'man 5 upsd.conf' and 'man 5 upsd.users'. This is a
configuration error.
My nut configuration is fine. To troubleshoot this, I had to label
each one of ACCESS-DENIED errors in the code, and with that, I
determined that the one involving tcp-wrappers was complaining. It
appears that local network communication is hard coded with
127.0.0.1. In my /etc/hosts.allow, I just have "ALL:localhost", with
no 127.0.0.1. TCP wrappers doesn't know or check that they're the
same, so I once I added it, it started working. :-) Presumably,
previous versions used localhost.
No, previous versions didn't use tcp-wrappers. That's why I pointed
you to 'man 8 upsd.conf' which has a paragraph ACCESS CONTROL that
tells you that we use tcp-wrappers. It's true that we only do a lookup
for the IP, not the hostname. Although it is common to include both
hostname and IP in hosts.allow, we probably should make a note to that
effect.
Best regards, Arjen
--
Please keep list traffic on the list
_______________________________________________
Nut-upsuser mailing list
[email protected]
http://lists.alioth.debian.org/mailman/listinfo/nut-upsuser
