[Nagiosplug-devel] [ nagiosplug-Bugs-998291 ] check_ping (1.3.1 and CVS) does not catch expired TTL errors
SourceForge.net
noreply at sourceforge.net
Sat Dec 4 12:30:05 CET 2004
Bugs item #998291, was opened at 2004-07-26 22:28
Message generated for change (Comment added) made by opensides
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=397597&aid=998291&group_id=29880
Category: Parsing problem
Group: Release (specify)
>Status: Pending
>Resolution: Fixed
Priority: 5
Submitted By: Georger Araujo (georger_br)
Assigned to: Benoit Mortier (opensides)
Summary: check_ping (1.3.1 and CVS) does not catch expired TTL errors
Initial Comment:
Guys,
I often get the following problem in my network:
[root at nagios nagios-plugins-1.3.1]# ping 10.83.0.242
PING 10.83.0.242 (10.83.0.242) 56(84) bytes of data.
>From 10.10.0.22: icmp_seq=0 Redirect Host(New nexthop:
10.10.0.242)
(... several lines of ICMP redirects ...)
>From 10.10.0.22 icmp_seq=0 Time to live exceeded
Nagios fails to figure out the host is down - in
reality, I have a routing loop and my routers are just
playing Pong.
Worse, as Nagios doesn't treat this as an error, the
ping goes on until check_plugin times out. Not good.
OK, you can say I should just append a -w to my command
file (when I do it, ping stops on the first "Time to
live exceeded" error) - but Nagios *should* figure this
out (that the host is down) and do its stuff.
The way check_plugin.c is now, Nagios will not play
nice with my network.
I just looked the latest check_ping.c in CVS and saw
that both the STDOUT/STDERR and -w issues have been
taken care of, but in order to attain as much accuracy
as possible please consider adding this to error_scan():
/* Georger's suggestion
if (strstr (buf, "Time to live exceeded"))
die (STATE_CRITICAL, _("PING CRITICAL - Time to live
exceeded (%s)"), addr);
*/
I'd love to use the CVS version, but it doesn't compile
on my box - I get an error stating that my_connect is
declared in two places (net_utils.h and my_sql.c, or
was it .h? Dunno, already deleted it), and I couldn't
figure out how to not compile check_mysql (or at least
compile just check_ping alone).
Long live Nagios.
Georger
----------------------------------------------------------------------
>Comment By: Benoit Mortier (opensides)
Date: 2004-12-04 21:29
Message:
Logged In: YES
user_id=388184
Ok,
i have applied to patch will be in cvs in the next hours
Thanks
----------------------------------------------------------------------
Comment By: Georger Araujo (georger_br)
Date: 2004-12-04 19:25
Message:
Logged In: YES
user_id=1087649
Yes,
I'm currently using the CVS version that was available when
I reported the bug, along with my hack.
My Nagios deployment works fine now.
----------------------------------------------------------------------
Comment By: Benoit Mortier (opensides)
Date: 2004-12-04 01:33
Message:
Logged In: YES
user_id=388184
Hi,
did you test the fix in 962495 and your little hack
any news ??
----------------------------------------------------------------------
Comment By: Georger Araujo (georger_br)
Date: 2004-07-26 22:33
Message:
Logged In: YES
user_id=1087649
Heh, just saw that the my_connect issue has been taken care
in bug report #962495. I'll try that in my network, along
with my little error_scan hack.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=397597&aid=998291&group_id=29880
More information about the Devel
mailing list