[Nagiosplug-devel] [ nagiosplug-Bugs-1867716 ] check_snmp invalid performance data

[SourceForge.net]

Bugs item #1867716, was opened at 2008-01-09 09:46
Message generated for change (Comment added) made by dermoth
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=397597&aid=1867716&group_id=29880

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: General plugin execution
Group: release-1.4.15
>Status: Closed
>Resolution: Fixed
Priority: 5
Private: No
Submitted By: Joerg Linge (pitchfork)
Assigned to: Thomas Guyot-Sionnest (dermoth)
Summary: check_snmp invalid performance data

Initial Comment:
check_snmp produces invalid performance data.

Version:
check_snmp v1859 (nagios-plugins 1.4.11)

Call:
check_snmp -H localhost -C public -o sysUpTime.0

Output:
SNMP OK - Timeticks: (49845368) 5 d | SNMPv2-MIB::sysUpTime.0=Timeticks: (49845368) 5 d

OS: Linux SuSE 2.6.11.4-21.17-bigsmp

Compiler:
gcc version 3.3.5 20050117 (prerelease) (SUSE Linux)

Kind regards
Jörg




----------------------------------------------------------------------

Comment By: Thomas Guyot-Sionnest (dermoth)
Date: 2010-03-30 22:40

Message:
This problem is now fixed in Git. Thank you for your report.

----------------------------------------------------------------------

Comment By: Thomas Guyot-Sionnest (dermoth)
Date: 2010-03-30 21:38

Message:
Reopened as this break half-working cases where although the method it
broken, then end result is somewhat OK. A proper fix should check the
string the same way as for getting the actual numeric value, and the string
can then be cleaned up using a two-way numeric-string conversion.

Example:
Pre-1.4.12:
$ ./check_snmp -H localhost -C public -o hrSWRunPerfMem.1
SNMP OK - 156 KBytes | HOST-RESOURCES-MIB::hrSWRunPerfMem.1=156 KBytes 

1.4.14:
$ ./check_snmp -H localhost -C public -o hrSWRunPerfMem.1
SNMP OK - 156 KBytes |  


v.1.4.12 has one valid perfdata string and a broken piece (what is
supposed to be the unit associated with the OID). Since this will normally
parse OK I'm counting that as a regression, although the fix was perfectly
valid in that it sanitized the performance data.

Will provide a better (hopefully) fix soon...

----------------------------------------------------------------------

Comment By: Matthias Eble (psychotrahe)
Date: 2008-07-09 17:55

Message:
Logged In: YES 
user_id=1694341
Originator: NO

this problem is now fixed in svn. thank you for your report.

----------------------------------------------------------------------

Comment By: Joerg Linge (pitchfork)
Date: 2008-01-12 06:07

Message:
Logged In: YES 
user_id=1353850
Originator: YES

sysUpTime.0 was just an example!

I think check_snmp should only provide performance data on interger
values.
Just to be sure the performance data ist valid.
 
Jörg

----------------------------------------------------------------------

Comment By: Thomas Guyot-Sionnest (dermoth)
Date: 2008-01-11 16:55

Message:
Logged In: YES 
user_id=375623
Originator: NO

This is the expected behavior; chhek_snmp doesn't know anything about
TimeTicks

Perhaps passing -Ot would help, but it would require different parsing and
I don't know how it would affect any the the many special data formats in
SNMP.

This would be better suited as a feature request. However it would be much
easier to just write a local check (run trough nrpe or check_by_ssh) that
parse uptime.

Also you should keep in mind that the SNMP uptime is the SNMP daemon
uptime, not the server uptime (in other words you're ok as long as you
don't restart the daemon).

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=397597&aid=1867716&group_id=29880