Submitted by koreth on 2003-01-24 08:48:22

check_by_ssh wasn't properly allocating memory for its
list of services, which lead to random odd behavior
especially with long lists of services.

Submitted by nobody on 2002-07-15 11:57:29

Is there any way check_http,https & ftp support a proxy,
so that you wan test proxy, as well as the site ?

Regards, Patrick

Submitted by bipsen on 2003-01-27 23:12:09

When checking the CPU load on a cisco pix firewall
using the command:

/usr/libexec/nagios/check_snmp -H hostaddress -
o .1.3.6.1.4.1.9.9.109.1.1.1.1.5.1 -C community -u '%
CPU Load' -w 80 -c 90

The result is:

SNMP WARNING - 0 % CPU Load

This is using net-snmp 5.0.7 (compiuled with ucd-snmp
compatability) and the plugins 1.3.0-beta2

snmp get on the oid gives 0 - it is just wierd, that the
plugin returns WARNING for this result...

Submitted by coltekin on 2002-11-20 12:48:53

check_tcp uses same buffer for sending and receiving for
send/expect strings. if remote service closes the
connection without
any data, check_tcp uses the same buffer for
comparison, and
if you expect to have same string, it fails to detect
the problem.

The small patch below works for me. It also ensures
that NULL string
is not passed to strip() (so strlen()) which core dumps
on my
(debian 3.0 with linux kernel 2.4.19, glibc 2.2.5) box
at least. This
is against nagiosplug-1.3-beta1.

Submitted by nobody on 2002-12-12 19:27:38

I have a problem with check_mysql plugin. It stops
always with a segfmenation fault on execution.
Here my configure options:
./configure
--prefix=/usr/local
--with-cgiurl=/nagios/cgi-bin
--with-nagios-user=nagios
--with-nagios-grp=nagios
--with-mysql=/usr/local/lib
--with-openssl=/usr/local/lib \

the mysql lib's under /usr/local/lib are:
...
libmysqlclient.a
-rwxr-xr-x 1 root root 652 Jul 12 15:01
libmysqlclient.la
-rw-r--r-- 1 root root 200402 Jul 12 15:01
libmysqlclient_r.a
-rwxr-xr-x 1 root root 678 Jul 12 15:01
libmysqlclient_r.la
...

I have testet it also with shared object files of
mysqllibs. The older version (beta1) of check_mysql
works but not the beta2 version!

My system is a linux 2.2.19 and the mysql version is a
3.23.47 - 3.23.51.

The ldd says:
ldd check_mysql
libz.so.1 => /usr/lib/libz.so.1 (0x2aac2000)
libc.so.6 => /lib/libc.so.6 (0x2aad1000)
/lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x2aaab000)

whats wrong?

Submitted by jtfalgout on 2002-07-03 21:29:39

Add the ability to query a specified port number to
check_snmp. This capability is available in snmpget
with the -p option.

Submitted by llow on 2002-06-13 02:18:38

I've added the -m option to check_snmp to allow you to
specify which mibs to use when doing the specified
snmpget. For large number of different types of
equipment being polled with different MIBs this is a
necessity to prevent wasted server resources.

NOTE: If you don't specify a mibs list, it will use the
default of UCD-SNMP which is the default mib list and
NOT the default of check_snmp which is all.

Submitted by purpletech on 2002-10-07 19:34:21

I patched check_dns.c to take another optional
parameter. If you specify "-a 12.34.56.78" then
(assuming everything else worked ok too) it compares
the returned address to the parameter. If they do not
match, it fails with CRITICAL.

Should it be a WARNING instead? I'm not familiar with
Nagios conventions as to what counts as serious enough
to be CRITICAL. I'm thinking, if it matters that the
service is running, use the normal check_dns; if you go
out of your way to test the returned address, then it's
important enough to you to count as a CRITICAL. But I
could be wrong.

I used -a and --expected-address as the parameter name.

Here's a template-style command for the new calling
style. I'm not sure what file in the project this
should go into. It seems that command.cfg needs to be
converted into the new template style...

check_dns_address!hostname-to-check!expected-ip-address

example: check_dns_address!myhost.foo.com!123.45.67.89

define command{
command_name check_dns_address
command_line $USER1$/check_dns -s $HOSTADDRESS$ -H $ARG1$ -a $ARG2$
}

A patch file is attached to this patch report.

Submitted by nobody on 2002-10-10 16:33:05

can we add an option to add checks for NT services?
That will be great because we can monitor anything in
NT if we can monitor the service. - srinivas A

Submitted by foresthus on 2002-07-26 14:02:48

Hi,

I´d like to check the PID of /usr/local/apache with is
started. The service is called httpd.

How can I do this?

and.. how do I use the plugins?

gero

Submitted by delorea on 2003-01-13 18:49:42

I found parsing error of the arguments in the
check_disk_smb.

Microsoft permit to have shares that ends with a $ but in
the parsing of the share parameter the $ is not
considered to be a valid character.

In the same way the user parameter don't parse
correctly because a user can be of this type:
domain\user. The \ char is not accepted.

I don't know if ther is already a patch.
If don't, I can modify this script to correct this problems.

Tnx
Ambrogio De Lorenzo

Submitted by jsellens on 2003-01-13 04:57:48

I want to check a web page that has a number of status
entries for
various servers - I'd like to check a regular
expression that
contains all those servers, but they are on multiple
lines.

check_http.c sets REG_NEWLINE which means that regular
expressions
may not span newlines. This means that, in my case, I
must check
the web page multiple times.

Here is a patch that adds an option -l / --linespan
that turns off
REG_NEWLINE, thereby allowing regular expressions to
span newlines,
and allowing me to check only once in my tests.

I've also removed some duplication of error messages if
HAVE_REGEX_H
is not defined, hopefully making the code slightly
clearer.

This is a patch against check_http.c version 1.13
(current in CVS).

Cheers!

John
[email protected]

Submitted by galactus on 2003-01-23 04:17:00

check_mysql.c as it appears in nagios-plugins-1.3.0-beta2.tar.gz wants the following patch to make it stop segfaulting

35d34
< int check_disk (int usp, int free_disk);
210,211c209,211
< if (is_intnonneg (argv[c]))

< db_port = atoi (argv[c++]);

  if (argc > c)
          if (is_intnonneg (argv[c]))
                  db_port = atoi (argv[c++]);

Submitted by cptmellow on 2002-04-15 21:26:04

Modified Perl script to search for redundant server
configurations like we use here. Added comments to code
describing the format our lmstat command produces so
as to describe how the new code works. The script
still handles the single license server instance as
well.

Submitted by tonvoon on 2002-12-11 16:24:01

Adds in support for Mac OS X / Darwin for ps command
syntax

Submitted by nobody on 2002-07-08 22:01:52

In the contrib subdirectory there is an example script
called check_ifstatus.pl used for checking Cisco
interface availability. The script works well but shows the
SNMP community on a failure. This is a security
concern as most people probably don't want their SNMP
strings known to anyone looking at the nagios web
interface.

Please find a simple patch for this below...

--- check_ifstatus.pl.orig Mon Jul 8 14:10:53 2002
+++ check_ifstatus.pl Mon Jul 8 14:11:01 2002
@@ -127,7 +127,7 @@
$answer=$session->error;
$session->close;
$state = 'CRITICAL';

•  print ("$state: $answer,$community,$snmpkey");
  

•  print ("$state: $answer");
  

  exit $ERRORS{$state};
  }

Submitted by nobody on 2002-08-07 11:50:58

Hi There, I haven't seen one, but I would love to have
a "check_https" plugin - allowing me to check SSL-
enabled web sites/servers

Cheers

Submitted by rcgreenw on 2002-07-19 23:12:02

Make check_smtp more rfc 821 compliant. It will send
HELO instead of MAIL. Postfix, and possibly
other mailers, don't like it when clients don't speak
valid smtp. This patch helps.

Submitted by nobody on 2002-08-07 22:25:33

Any of you got purify or some other memory debugging tool? you
need to run it on check_snmp which suffers from some massive
problems.

as it is, there are tons of errors: Free Memory
Read, Write and even memory being free several times. These
eventually lead to realloc returning NULL because it's called on
memory not allocated and check_snmp aborts
violently.

commenting out the free(ptr) in ssprintf fixes the
problems which in turns allows check_snmp to work properly.
However, it now leaks memory.

this is again for checking 2
OIDs at the same time, and using -c and/or -w

Submitted by nobody on 2002-10-24 21:11:17

The check_snmp module can not be made to return a value
of higher than 'WARNING' regardless of the options
specified.

Even a simplistic request, ie

'check_snmp -H -C -o
system.sysUpTime.0 results in the following:

SNMP WARNING - Timeticks: (1516539759) 187 days,
2:23:17.59

build is 1.3 beta1, OS is debian/woody.

Submitted by ghelleks on 2002-07-12 07:17:42

• option to show only devices in error
• option to exclude a single device (useful for CDROMs)

Submitted by nobody on 2002-05-23 21:35:04

None of the plugins available via CVS (nor the "older"
Netsaint version of them) seems to pass the
$PERFDATA$ macro to Nagios, make it quite difficult to
record that data for further analysis without hammering
down the plugins code.
I'm not a C programmer, so my own "code hammering"
is ugly ... well, very ugly :)

C'mon guys, ... appending " |
name_of_the_checked_vars=%s" to the output of the
plugins cannot be THAT hard :-)

Submitted by nobody on 2002-06-13 22:26:34

Need to send a after the command per RFC 821... Thanks to Tom Bertelson and Benjamin
Madsen for pointing out the issue. Patch has been applied to CVS.

--- check_smtp.c Thu Jun 13 16:22:42 2002
+++ ../../nagiosplug/plugins/check_smtp.c Sun Jun 9 23:24:06 2002
@@ -48,7 +48,7 @@

• even if syntax is illegal (MAIL requires a FROM:<...> argument)
• You can disable sending DUMMYCMD by undefining SMTP_USE_DUMMYCMD.
  */
  -#define SMTP_DUMMYCMD "MAIL\r\n"
  +#define SMTP_DUMMYCMD "MAIL\n"
  #define SMTP_USE_DUMMYCMD 1
  #define SMTP_QUIT "QUIT\n"

Submitted by bhat on 2002-06-06 04:36:16

check_rbl uses dig to check if an IP address is listed
on the specified
RBL. check_rbl is based on the check_dig code, and is
thus under the
GPL.

check_rbl is coded as a plugin for netsaint. I hope it
can be incorporated
into nagios as well.

Tim Bell.
[email protected]

Submitted by mgrape on 2002-10-04 15:09:02

When I run check_smtp against our Exchange server it
first outputs the "SMTP OK - 0 second response time"
line, then it hangs for the timeout period and then it fails
with a "Socket timeout after seconds"
message.

It works fine against sendmail on a solaris 7 machine.

plugins 1.3 beta1
Solaris 7
gcc 2.95.2

If you need any tests done, feel free to contact me.

Submitted by wme on 2002-08-28 19:32:58

intra02# make all
Making all in plugins
"Makefile", line 760: Need an operator
make: fatal errors encountered -- cannot continue
*** Error code 1

Stop in /tmp/nag/nagiosplug-1.3-beta1.

Don´t know where to search in a Makefile of 387 Lines.

nagios itself made no problems (until now :)
Any ideas?? Thanks in advance

wme

Submitted by nobody on 2003-01-06 12:15:28

  Check_ldap segfaults if argc>=2 and the -H or -b options are not supplied.  This is because 

validate_arguments tries to reference elements of the ld_base and ld_host arrays that might not exist.
This appears to be just a typo, and I fixed it. Hopefully. Also, the syntax help message that the program
spits out doesn't explain that the -c and -w options are required. I fixed up the syntax help a bit to
explain that. I also added a slightly more specific explanation for some errors in the command line.

Andrew McDavid
[email protected]
(remove spamfree)

Submitted by ricozz on 2002-10-28 19:53:11

It seems to be usefull to reverse -r|-R|-s switch result.

For example, i have a monitor.php script which test a
webserver
(connection to a database, php engine on etc..) and
return tests OK if it's
good or ERR if doesn't

I want to use
./check_http -I somehost --reverse -r '[A-Z]+ : ERR' -u
/monitor.php -v

Page is 424 characters
STATUS: HTTP/1.1 200 OK
**** HEADER ****
Date: Sun, 27 Oct 2002 10:58:59 GMT
Server: Apache/1.3.26 (Unix) Debian GNU/Linux PHP/4.2.2
X-Powered-By: PHP/4.2.2
Connection: close
Content-Type: text/plain; charset=iso-8859-1
**** CONTENT ****
BASEC : ERR
==>2002: Can't connect to local MySQL server through
socket...
BASES : ERR
==>2002: Can't connect to local MySQL server through
socket ...
PHP : OK

HTTP CRITICAL: pattern found

Submitted by dts12 on 2002-07-23 10:54:03

The plugins 'check_snmp' and 'check_hpjd' use the
command-line tool 'snmpget' from the {UCD,Net}-SNMP
suite. Up to v4, this tool could accept an SNMPv1
community string as a positional parameter, or using
the explicit command option '-c'. Starting from v5,
this tool only uses the option '-c'. Attempting
to use the positional form (as here) will fail.

The attached patch tweaks these two plugins to use
the -c option instead.

Dave

Submitted by lachlancs on 2002-09-05 03:13:19

check_dns_random.pl doesn't use the exit value from the
check_dns command. Attached is a diff to exit from
check_dns_random.pl with the exit value of the
check_dns command.

Hope you find it useful!

Lachlan

Submitted by nobody on 2002-07-26 16:09:08

Sorry if this appears twice... I submitted it once, but it
seems to have disappeared...

The -g option wasn't actually put in the "list of known
options" for argument processing in check_nntp,
causing it to usage-die if used.

Submitted by nobody on 2002-07-09 21:04:03

In the contrib directory there is a script called
checkciscotemp.pl. This script works well but I would
recommend the following changes.

• Change the name from checkciscotemp.pl to
  check_ciscotemp.pl to better fit nagios-plugins naming
  conventions
• Add the following patch to clean up a Perl issue
  regarding implicit splits. The script works with this patch
  but it does involve removing error checking code. I am
  not familiar enough with Perl to make a "proper" patch
  for this problem but the patch shown below works well in
  a production environment.

--- /usr/lib/nagios/plugins/contrib/checkciscotemp.pl
Sun Jul 7 22:26:35 2002
+++ checkciscotemp.pl Tue Jul 9 13:57:42 2002
@@ -81,15 +81,11 @@
}

if (defined($critical_vals)) {

• die "Cannot Parse Critical Thresholds\n"

•   unless (split(/:/,$critical_vals)>=2);
  

• ($inlet_thresh,$outlet_thresh) = @_
• ($inlet_thresh,$outlet_thresh) = @_;
  }
  die unless(defined($inlet_thresh) && defined
  ($outlet_thresh));

if (defined($warning_vals)) {

• die "Cannot Parse Critical Thresholds\n"

•   unless (split(/:/,$warning_vals)>=2);
  

  ($inlet_warn,$outlet_warn) = @_;
  }else{
  $inlet_warn=$inlet_thresh;

Submitted by fpehla on 2002-11-27 16:35:46

Hello,

after switching to asprintf, check_procs prints (NULL) in
status lines on Solaris.

Please find enclosed an appropriate patch to fix that.

Best regards,
Fabian

Submitted by apollock on 2002-07-04 03:42:26

Hi,

Using (nagios-plugins 1.3.0-alpha1) 1.1.1.1

I'm doing a ./check_http -H www.google.com and
it's timing out. Yet it works if I am to telnet
normally (it is transparently proxied). I've attached
part of an strace, where you can clearly see the
HTTP response returned from the GET request.

Any ideas why this is timing out?

Submitted by sghosh on 2002-04-07 18:04:27

I tried compiling nagios w/ embedded perl and without caching, but two of the
netsaint plugins I use gave me this error:
UNKNOWN: Unable to resolve destination address '-c'

Specifically, it was check_ifstatus.pl and check_maxwanstate.pl. Also,
check_ntp just returned critical, as if it couldn't find the host I
specified.

• -Doug

Submitted by mkingsbury on 2002-11-16 08:41:03

The -p option on check_disk (CVS version 1.4) is
broken. Given an option -p "/ /boot", where / is /dev/md0
and /boot is /dev/hda1, I get the output:
"DISK OK - [27018 kB (61%) free on /dev/hda1]
[27018kB (61%) free on /dev/hda1].

where the /boot partion is displayed twice rather
than /dev/md0.

The version in the v1.3 Beta1 tarball does not have this
problem.

-mike

Submitted by ulab on 2002-09-03 13:53:58

nagiosplug-1.3-beta1

If I use check_smtp, postfix sends a message to postmaster:

Out: 220 foghorn.bartels-schoene.de ESMTP Postfix
In: MAIL
Out: 501 Syntax: MAIL FROM:

I wonder why MAIL is sent anyway? If you really want to
check the response time you'd better use real
SMTP-Commands ;-)

 Balu

Submitted by briansaf on 2002-08-09 18:15:53

First time installiing 'Netsaint/Nagios' but, heard alot of
good things about it.

The base installed just fine. Running Solaris 8 and
using Cygnus gcc.

Installing nagiosplug-1.3.beta1 using the following;

./configure --prefix=/home/nagios/nagios --with-nagios-
group=std

This appears to have worked OK.

Next ran 'make all' and received the following error
message:

friday# make all

Making all in plugins
mksh: Fatal error in reader: = missing from replacement
macro reference
Current working directory /home/nagios/nagiosplug-1.3-
beta1/plugins
*** Error code 1
make: Fatal error: Command failed for target `all-
recursive'

friday# pwd

/home/nagios/nagiosplug-1.3-beta1

Submitted by nobody on 2002-07-08 21:25:46

In the contrib subdirectory there is an example script
called check_ifstatus.pl used for checking Cisco
interface availability. The script works well but shows the
SNMP community on a failure. This is a security
concern as most people probably don't want their SNMP
strings known to anyone looking at the nagios web
interface.

Please find a simple patch for this below...

--- check_ifstatus.pl.orig Mon Jul 8 14:10:53 2002
+++ check_ifstatus.pl Mon Jul 8 14:11:01 2002
@@ -127,7 +127,7 @@
$answer=$session->error;
$session->close;
$state = 'CRITICAL';

•  print ("$state: $answer,$community,$snmpkey");
  

•  print ("$state: $answer");
  

  exit $ERRORS{$state};
  }

Submitted by loki77 on 2002-07-10 22:07:55

I am having the following problems trying to build
nagiosplug-1.3-beta1.tar.gz on Solaris 8. I'm using gcc
and gnu make. Here's the error I get when I attempt to
do a make all:

$ make -v
GNU Make version 3.79.1, by Richard Stallman and
Roland McGrath.
Built for sparc-sun-solaris2.8
Copyright (C) 1988, 89, 90, 91, 92, 93, 94, 95, 96, 97,
98, 99, 2000
Free Software Foundation, Inc.
This is free software; see the source for copying
conditions.
There is NO warranty; not even for
MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.

    Report bugs to <[email protected]>.

$ make
cd . &&
/bin/sh: syntax error at line 1: `end of file' unexpected
make: *** [aclocal.m4] Error 2

Submitted by tonvoon on 2002-12-11 16:35:33

Fixes:
check_procs -p 1 (bus error)
check_procs -p 1 -u 0 (wrong output)

Submitted by klgeddie on 2002-12-17 17:38:00

The check_real plugin does not work as distributed. The
code attempts to check the server assigned to the
host_name variable, but the host_name variable is never
assigned a value.

The following patch will fix the code:

*** check_real.c Mon Dec 16 13:21:50 2002
--- check_real.c.orig Thu Nov 14 20:04:51 2002

*** 400,413 ****
int
validate_arguments (void)
{

- int rc = OK;

•   if ( server_address  &&  *server_address )
  

•   {
  

•           host_name = server_address;
  

•   } else {
  

•           rc = ERROR;
  

•   }
  return OK;
  

  }

--- 400,405 ----

Submitted by nobody on 2003-01-10 14:15:56

I try the command "check_snmp -H testhost -o
system.sysUpTime.0" and get a WARNING state.

Submitted by bkw on 2002-11-06 13:19:11

It would be really nice to have a feature like
check_http's --certificate for check_smtp also. Issue a
STARTTLS and analyse the given cert.

Any chance to see something like that soon? Or sould I
try to roll my own?

cheers,
bkw

Submitted by nobody on 2002-09-25 07:14:53

Hi,

We are behind a proxy server that requires
authentication. We have to type username/password to
access the internet and because of that limitation its not
possible to use check_http. It would be great if you can
add proxy authentication option in the check_http.

Thanks,

Submitted by tonvoon on 2002-12-09 18:07:46

This is a patch to the check_oracle script so there is a --
login check. This is better than a --db and a --tns check
because it makes a connection to the database, uses a
dummy id, and then checks that there was a ORA-01017:
invalid username/password.

For example, this picks up a ORA-01033: ORACLE
initialization or shutdown in progress which the standard
check_oracle tests will miss. Will also get Oracle's ORA-
error message in the returned output.

It does not require a valid userid/password.

Also moved the --help and --version higher up so that these
messages will be seen before the script complains about an
the ORACLE_HOME variable.

Submitted by rosenauer on 2002-07-09 12:06:26

I don't know if it is intention but I wanted to be able
to use SNMPv3
with check_snmp and made the following patch.

Submitted by tonvoon on 2002-12-11 16:39:06

Fixes:
check_vsz (no error reported)
check_vsz -A (two invalid option messages)
check_vsz 1000 (bus error)
check_vsz 100000 (always gives critical error - now highest
alert is warn)
check_vsz -C init (no error reported)
check_vsz -c 100000 (always gives warning)

Submitted by agraaff on 2002-11-05 10:57:09

I have some trouble retriving counters values from the
NT4 and W2000 performance monitor.
I have found that when you try to retrive a counter which
contains a '/' in the name you will get a number back
that makes no sense.
I tryed to do this with several counters containing
forward slashes and found the same problems.

Example:
./check_nt -H 10.11.250.10 -v COUNTER -
l "\Memory\Page Faults/sec","%.3f"

because of the '/' in "Page Faults/sec" something goes
wrong with the query. i do get a value back but it is not
the same value i see when i look at perfmon in
Windows.

What is the value i get here and how can i get the value i
requested?

Submitted by nobody on 2002-08-07 21:36:18

In particular, when trying to use check_snmp with two OIDs, the
command line used to invoke snmpget is truncated in the middle of
the second OID resulting in a failure.

i bumped TXTBLK to
1024 as a quick fix.