Zext_nagios.sh

[wikisb]

hi all!
i'm trying to put parameters in a item to monitorize a service but... i don't know how the sintax works of zext_nagios.sh.
how do you add parameters??
for example to do a check_ping you have to put 3 parameters (hostname or ip, warning thereshold, critical thershold) for example with:
"HOSTNAME -w 100,60% -c 100,60%"
how it would be??????
zext_nagios[ping {HOSTNAME}.... ]
i have tried to put this parameters in diferents forms but i haven't been able to solve this issue. i would be very glad if someone helps me.
it only gives me value 2 or 3 for critical or unknown
thanks !!

[JBo]

Hi,

zext_nagios.sh[ping {HOSTNAME} -w 100,60% -c 100,60%]

works.

I have attached a template that defines it and associated performance items (ping_rta, ping_pl).

Regards,
JBo

[wikisb]

hi JBo!!!!
thanks again for the post!! but i don't know what i'm doing wrong because i still get values 1 to 3 when i try to monitorize services with zext_nagios.sh that needs more parameters than the HOSTNAME.
if i do
zext_nagios.sh[ssh {HOSTNAME}] it gives me 0 when ssh is running & when i stop ssh it gives me 2, value for critical. so it works OK like check_pop and the others checks that only need the HOSTNAME as parameter.
But when i try to monitorize postgres, ping and such other services that need more than one parameter it doesn't works, with check_ping i have put the code you posted "zext_nagios.sh[ping {HOSTNAME} -w 100,60% -c 100,60%]" but it gives me value 3 when it should be 0. i also imported your template and linked it to a host and it gives me value 3 for UNKNOWN. if i run ./check_ping HOSTNAME -w 100,60% -c 100,60% in console it gives me OK
with check_postgres it happens the same, if i put "zext_nagios.sh[pgsql {HOSTNAME} -l nobody -p nobody]" it gives me value 3 for UNKNOWN but i know that if i run in the terminal "./check_pgsql 192.168.1.16 -l nobody -p nobody" it gives me OK.
i don't use zabbix_agent but it shouldn't matter to do external checks.
thanks one again and i hope we can solve this in brief.
and if it wheren't for you JBo i wouldn't be monitoring any services at all. so special thanks for you

[JBo]

Hi,

I have tested

 zext_nagios.sh[ping {HOSTNAME} -w 100,60% -c 100,60%]

on my Zabbix installation and it is working fine.

I don't know what is wrong on your setup.
You can try to edit zext_nagios.sh and set

DEBUG=1

at the beginning of the script.

It will send a lot of traces to /tmp/zext_nagios.log, it may help you find what is wrong.

Regards,
JBo

[wikisb]

hi
i just put debbug=1 and the zext_nagios.log prints this:
the nagios plugin path is the one i edited from zabbix_Server.conf

+ NAGIOS_PLUGIN_PATH=/usr/lib/nagios/plugins
+ export LANG=C
+ LANG=C
+ ip=182.10.1.16
+ plugin=ping
+ host=fsystem
+ shift 3
+ '[' -w = -p ']'
+ prefix=ping
++ /usr/lib/nagios/plugins/check_ping -H 172.20.1.16 -w 100,60% -c 100,60%

i don't know if there is something not expected in this zext_nagios.log...

[JBo]

Hi,

This is the trace I get (host is www.google.com and I have set critical level to 200,90%):

+ NAGIOS_PLUGIN_PATH=/usr/lib/nagios/plugins
+ export LANG=C
+ ip=www.google.com
+ plugin=ping
+ host=www.google.com
+ shift 3
+ [ -w = -p ]
+ prefix=ping
+ /usr/lib/nagios/plugins/check_ping -H www.google.com -w 100,60% -c 200,90%
+ res=PING OK - Packet loss = 0%, RTA = 31.58 ms|rta=31.584999ms;100.000000;200.000000;0.000000 pl=0%;60;90;0
+ code=0
+ sed -n s/.*|//p
+ echo PING OK - Packet loss = 0%, RTA = 31.58 ms|rta=31.584999ms;100.000000;200.000000;0.000000 pl=0%;60;90;0
+ res=rta=31.584999ms;100.000000;200.000000;0.000000 pl=0%;60;90;0
+ zabbix_sender --zabbix-server 127.0.0.1 -i -
+ echo rta=31.584999ms;100.000000;200.000000;0.000000
+ sed -e s/=.*$//
+ param=rta
+ sed -e s/.*=// -e s/[^0-9\.].*//gi
+ echo rta=31.584999ms;100.000000;200.000000;0.000000
+ value=31.584999
+ echo www.google.com ping_rta 31.584999
+ sed -e s/=.*$//
+ echo pl=0%;60;90;0
+ param=pl
+ sed -e s/.*=// -e s/[^0-9\.].*//gi
+ echo pl=0%;60;90;0
+ value=0
+ echo www.google.com ping_pl 0
+ echo 0

If your trace stops at
/usr/lib/nagios/plugins/check_ping -H 172.20.1.16 -w 100,60% -c 100,60%
it seems that check_ping is blocked.

Does
/usr/lib/nagios/plugins/check_ping -H 172.20.1.16 -w 100,60% -c 100,60%
return something when run from the command line ?

JBo

[wikisb]

hi JBo
the first time i realized i wasn't getting the wright values for ping, postgres, etc.. i runned them with the terminal "./check_ping 182.10.1.16 -w 100,60% -c 100,60%" or "./check_pqsql 182.10.1.16 -l nobody -p nobody" and both of them returned OK with several info... but putting it in zabbix it says item not supported...
i don't know what to do.... it seems wierd that everything else works except the zext_nagios.sh with more than one parameter. i have monitorized lots of items i also have several triggers that work fine, graphs ,etc...
I WANT AN OUTPUT of zext_nagios.log as yours!!!!
thanks for helping me!!

[JBo]

I don't have any clue on what is going on.

Just a little thing:
zext_nagios.sh uses:
/usr/lib/nagios/plugins/check_ping -H 172.20.1.16 -w 100,60% -c 100,60%

You test it with:
./check_ping -H 172.20.1.16 -w 100,60% -c 100,60%

What do you get with the exact same command as zext_nagios.sh (including /usr/lib/nagios/plugins path) ?

JBo

[wikisb]

hi JBo

[root@fsystem ~]# /etc/zabbix/externalscripts/check_ping -H 182.10.1.16 -w 100,60% -c 100,60%
PING OK - Packet loss = 0%, RTA = 0.01 ms|rta=0.015000ms;100.000000;100.000000;0.000000 pl=0%;60;60;0

this is the same output as doing it without all the path
i changed the path of the scripts in zabbix_server.conf and zext_nagios.sh so all the scripts are in /etc/zabbix/externalscripts

[JBo]

hi JBo

[root@fsystem ~]# /etc/zabbix/externalscripts/check_ping -H 182.10.1.16 -w 100,60% -c 100,60%
PING OK - Packet loss = 0%, RTA = 0.01 ms|rta=0.015000ms;100.000000;100.000000;0.000000 pl=0%;60;60;0

this is the same output as doing it without all the path
i changed the path of the scripts in zabbix_server.conf and zext_nagios.sh so all the scripts are in /etc/zabbix/externalscripts

OK but /etc/zabbix/externalscripts is not the path that appears in your debug log.
You have:

+ prefix=ping
++[B] /usr/lib/nagios/plugins[/B]/check_ping -H 172.20.1.16 -w 100,60% -c 100,60%

Sorry to pin point what may be obvious to you but, since I have no explanation, I am tracking everything that looks strange

JBo

[wikisb]

hi JBo
i have tried lots of things but i haven't been able to solve the issue so finally about an hour ago i changed the paths of zabbix_server.conf where fetches the scripts and the path of zext_nagios.sh where fetches the script to /etc/zabbix/externalscripts. so now in zext_nagios.log i see

NAGIOS_PLUGIN_PATH=/etc/zabbix/externalscripts

but i get the same sh*t as before...
i'm sorry that you have put lot of your time to try to help me
thanks!!

[wikisb]

hi JBo!
still i didn't solved all the problem, but with some changes in permissions and giving shell to user nobody (it didn't had one till this morning) now i can monitorize all the services except "PING" but i don't mind if only doesn't works service ping, i will monitorize it with another thing instead using nagios scripts. but service postgres works fine like all the others.

the last thing i tried with check_ping was:

res='strace $NAGIOS_PLUGIN_PATH/check_$plugin -H $ip $*'

the last line returned was

write(1, "Ping OK".....) =-1 broken Pipe

thanks for your cool script

[JBo]

Hi,

Happy to know you could make some progress

According to strace, check_ping is working but it cannot write to its standard output.

You may try to replace:

res=`$NAGIOS_PLUGIN_PATH/check_$plugin -H $ip $*`
code=$?
res=`echo $res | sed -n 's/.*|//p'`

with:

$NAGIOS_PLUGIN_PATH/check_$plugin -H $ip $* > /tmp/zext_nagios.$$.out
code=$?
res=`cat /tmp/zext_nagios.$$.out | sed -n 's/.*|//p'`
rm /tmp/zext_nagios.$$.out

It may give a clue about what is going on.

Regards,
JBo

[wikisb]

Hi JBo!!
i tried what you posted yesterday afternoon. i have put a strace in the first line to be modified.

strace $NAGIOS_PLUGIN_PATH/check_$plugin -H $ip $* > /tmp/zext_nagios.$$.out

and know it doesn't says Pipe broken but still it gets stuck in this line never passing to the next lines. it creates the zext_nagios.$$.out that contains the response of PING with OK, rta... but it never removes it.
the last lines of the output of strace are the following:

read(3, "PING 182.10.1.16 (182.10.1.16) 5"..., 4096) = 112
read(3, "64 bytes from 182.10.1.16: icmp_"..., 4096) = 59
read(3, "64 bytes from 182.10.1.16: icmp_"..., 4096) = 59
read(3, "64 bytes from 182.10.1.16: icmp_"..., 4096) = 59
read(3, "64 bytes from 182.10.1.16: icmp_"..., 4096) = 60
read(3, "--- 182.10.1.16 ping statistics "..., 4096) = 149
--- SIGCHLD (Child exited) @ 0 (0) ---
read(3, "", 4096)                       = 0
read(5, "", 4096)                       = 0
close(5)                                = 0
munmap(0x7f21c81af000, 4096)            = 0
close(3)                                = 0
munmap(0x7f21c81b0000, 4096)            = 0
wait4(8104, [{WIFEXITED(s) && WEXITSTATUS(s) == 0}], 0, NULL) = 8104
fstat(1, {st_mode=S_IFREG|0664, st_size=0, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f21c81b0000
write(1, "PING OK - Packet loss = 0%, RTA "..., 102) = 102
exit_group(0)                           = ?

thanks!!

[JBo]

Hi,

It seems that zext_nagios.sh script is killed as soon as check_ping terminates.

It is possible that it is killed by Zabbix because of a timeout.
You can check Timeout parameter in zabbix_server.conf

Timeout=30

I set the timeout to its maximum value (30s) because I have a few checks that may take several seconds.

Regards,
JBo