Quantcast
Channel: Monitoring-Portal Feed
Viewing all 1338 articles
Browse latest View live

OMD 1.30 - Sonderzeichen escapen

$
0
0
Hallo,

ich habe mir heute morgen OMD geupdatet auf 1.30. Leider funktioniert seither ein classic active nagios check nicht mehr den ich in WATO konfiguriert habe.

so steht es schon seit fast einem jahr in WATO und hat immer funktioniert.

Quellcode

1
/opt/omd/sites/intern/lib/nagios/plugins/check_nut_plus -d apc-3000@xxx.xxx.xxx.xxx -v 'ups.status=c\!=OL'


auf der konsole funktioniert der check auch, allerdings so:

Quellcode

1
/opt/omd/sites/intern/lib/nagios/plugins/check_nut_plus -d apc-3000@xxx.xxx.xxx.xxx -v 'ups.status=c!=OL'


So, wie bekomme ich den unter WATO wieder zum Laufen? Ohne Escape gehts nicht, mit doppeltem Escape auch nicht :(

gruesse

Geschweifte Klammern in Syntax

$
0
0
Hallo,

mit dern { } werden entsprechend die einzelnen Definitionen markiert. Wie muss ich die denn schreiben, wenn ich welche in einem Check benutzen möchte.

In den NRPE Argumenten für den NSClient++ werden diese z. B. für die top-syntax benötigt. Normales "escapen" mit \{ \} habe ich natürlich bereits probiert.

Grüße

Templates über Service-Name anstatt check_command-Name ansprechen

$
0
0
Hi zusammen,

wie in der Überschrift gesagt, will ich bei Icinga2 Custom Templates/Graphen nutzen. Jedoch habe ich das Problem, dass ich in der services.cfg für jeden Check den selben check_command (nrpe) nutze. Für die Service-Definition nutze ich die Variable vars.nrpe_command. Jedoch werden die pnp4nagios-Templates ja über den check_command gesucht.

So sehen die Configs für einen Check aus:

===========================================================================
commands.cfg:

object CheckCommand "nrpe" {
import "plugin-check-command"
import "ipv4-or-ipv6"

command = [ PluginDir + "/check_nrpe" ]

arguments = {
"-H" = "$nrpe_address$"
"-p" = "$nrpe_port$"
"-c" = "$nrpe_command$"
"-n" = {
set_if = "$nrpe_no_ssl$"
description = "Do not use SSL."
}
"-u" = {
set_if = "$nrpe_timeout_unknown$"
description = "Make socket timeouts return an UNKNOWN state instead of CRITICAL"
}
"-t" = "$nrpe_timeout$"
"-a" = {
value = "$nrpe_arguments$"
repeat_key = false
order = 1
}
}

vars.nrpe_address = "$check_address$"
vars.nrpe_no_ssl = false
vars.nrpe_timeout_unknown = false

timeout = 5m
}
==============================================================

services.cfg:


apply Service "load" {

import "generic-service"
check_command = "nrpe"
vars.nrpe_command = "check_load"

assign where "services-server" in host.groups
assign where "linux-server-stg" in host.groups
}

============================================================

templates.cfg:


template Service "generic-service" {
max_check_attempts = 5
check_interval = 1m
retry_interval = 30s
import "pnp-svc"
}


template Host "pnp-hst" {
action_url = "/pnp4nagios/graph?host=$HOSTNAME$' class='tips' rel='/pnp4nagios/popup?host=$HOSTNAME$&srv=_HOST_"
}
template Service "pnp-svc" {
action_url = "/pnp4nagios/graph?host=$HOSTNAME$&srv=$SERVICEDESC$' class='tips' rel='/pnp4nagios/popup?host=$HOSTNAME$&srv=$SERVICEDESC$"
}

=========================================================================================

Gibt es denn nun irgendeine Möglichkeit, dass pnp4nagios nicht nach dem check_command sondern nach dem service-command sucht? Denn aktuell würde ich einen Custom-Graphen für jeden Check erhalten, die nrpe als check_command nutzen, also für alle den selben 8| :D
Danke für Euere Hilfe!




Gruß Flo

Notification über Threema Gateway API

$
0
0
Hallo zusammen,
wir haben OMD 1.2 auf einem Ubuntu Server 14.4 und wollen nun die Notification über das threema Gateway laufen lassen.

Wer hat Erfahrung damit?
Was habt ihr gemacht?
Gibt es schon ein passendes Plugin?



die Leseratte

cairo update on RHEL 5.11 Tikange

$
0
0
Hi,

I am getting this error when I try to run yum update:

Skipping security plugin, no data
--> Running transaction check
---> Package cairo.i386 0:1.8.8-1 set to be updated
--> Processing Dependency: libpixman-1.so.0 for package: cairo
--> Finished Dependency Resolution
cairo-1.8.8-1.i386 from labs_consol_stable has depsolving problems
--> Missing Dependency: libpixman-1.so.0 is needed by package cairo-1.8.8-1.i386 (labs_consol_stable)
Error: Missing Dependency: libpixman-1.so.0 is needed by package cairo-1.8.8-1.i386 (labs_consol_stable)
You could try using --skip-broken to work around the problem
You could try running: package-cleanup --problems
package-cleanup --dupes
rpm -Va --nofiles --nodigest
The program package-cleanup is found in the yum-utils package.

I found no newer package than this: qpixman-0.13.3-4.el5.i386
[101 KiB]

Can you help?

Thanks,

klajsoh2

Missing Hostgroups since update

$
0
0
Hello,

I recently upgraded my Icingaweb to version 2.0.0 and since this upgrade I am missing a lot of custom hostgroups even though I changed nothing, they worked fine before.
Below are screenshots of my current hostgroups in Icingaweb and a part of my groups.conf and a host configuration file which shows how I created these custom groups.





As you can see I have multiple groups defining the OS of each server, as well a "Servers" hostgroup which combines Windows and Linux. This always worked prior to the update, has something changed? Also I know I don't need the commented text in my hosts file, I just have it there as a reminder for people who aren't experienced or if I forget something.

Very high load in cluster setup

$
0
0
Hi,

I've got a new Icinga2 setup running, large number of hosts (715 hosts) and services (2657)
I'm currently running them on VM servers with intermediate resources (4 4Ghz CPUs, 4Gb ram) for testing.

Configuration of zones.conf

Zitat

/etc/icinga2/zones.conf

object Endpoint "monitor-one" {
host = "monitor-one.localdomain.net"
}

object Endpoint "monitor-two" {
host = "monitor-two.localdomain.net"
port = "5665"
}

object Zone "master" {
endpoints = [ "monitor-one", "monitor-two" ]
}

object Zone "global-templates" {
global = true
}


I start up the first server (This is what I've classed as 'primary' and has the configuration stored in the /etc/icinga2/zones.d/) and load initially goes very high as it starts with no cached state, etc, and runs as many parallel checks as it can. Load goes up over 50 on this box and slows down quite considerably (is this right? This high load, it seems to overload the machine trying to catch up).
However, leave this for 30-40 minutes and load will slowly ease and checks go into their normal sequence.

I wait for this to happen, then go to start the second icinga2 up on monitor-two, and this seems to cause an overload on both machines. The second instance starts up, pulls the config from the first box, config gets checked OK and starts up. This machine then starts to run its checks, load goes up and over 150 on this machine, but at the same time it goes up on the first instance too, over 100, and seems to stay that high. I've had it running for over 30 minutes and the load doesn't seem to change, just stays very high.

I've confirmed the cluster sync, authentication, etc, is working correctly, etc.

Checking the process table I can see that on both machines it is primarily the check commands filling up the process table, and also mysqld keeps jumping up (These machines only run icinga2, no other tasks).

Can anyone advise on what could be going wrong, or where I've gone wrong with the cluster? Or what additional information I need to check into to diagnose further.

Thanks!

Automatic package configuration for deployment of Icinga2 windows agent

$
0
0
Hi,

I have setup the agent on some windows hosts, and they report correctly their services to the master.

My question is : can I make this a little more automated ? not a problem to deploy in my current environment as I have only a few windows hosts to manage, but if I'm going to deploy icinga2 in bigger infrastructures, it's going to be a problem.

Steps I'd like to automate :

- Generate the ticket hash on the master, using the FQDN.
- Create a package configuration file (XML for an MSI of batch file with parameters for the installer)

I'll probably deploy the packages though GPO or something like that.

Anyone has done that ?
Thank you for any help.

check_mk und oracle_health zusammen verwenden

$
0
0
Hallo Zusammen,

ich verzweifle langsam aber sicher...

Habe es nun endlich geschafft oracle_health in Verbindung mit OMD und perl zum laufen zu bekommen.

Nun wollte ich einige Tests einrichten uhnd scheiter gerade an der konfiguration.

Ich habe mich für check_mk entscheiden und möchte nun das Plugin check_oracle_health einbinden. Hierzu möchte ich gerne eigene Variablen definieren und befüllen und nätürlich die Tests ausführen. Des Weiteren würde ich gerne das Feld Aliase für den Parameter --connect verwenden. Leider raffe ich nicht, wie man Plugins in check_mk verwenden kann, welche nicht zum Standard gehören.

Aktuell habe ich in der main.mk folgendes drin:

Quellcode

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
#all_hosts = [ ]
 # Put your host names here
 # all_hosts = [ 'localhost' ]
 all_hosts = [
  'DB1|oracle|'
  ,'DB2|oracle|'
  ,'DB3|oracle|'
 ]  
 define_hostgroups = True
  
 host_groups = [
 ( 'ORACLE', ['oracle'], ALL_HOSTS ),
 ]
  
 _ORACLEUSER = "oracleuser"
 _ORACLEPASSWORD = "password"
  
 #Legacy Checks
 extra_nagios_conf += r"""
  
 define command {
  command_name check_oracle_tnsping
  command_line $USER1$/check_oracle_health --connect "$ALIAS$" --user "$_ORACLEUSER$" --password "$_ORACLEPASSWORD$" --mode tnsping
 }
  
  
 """
  
 legacy_checks = [
  ( ( "check_oracle_tnsping!", "TNS-PING", True ) ),
 ]


Was mache ich nur falsch???

Würde mich über einen hilfreichen Hinweis sehr freuen. Auch über eine genauere erklärung zum Vorgehen. :-)

Vielen Dank im Voraus...

VG micha_bln

check_http mit string

$
0
0
Hallo zusammen!

Ich würde gerne eine Webseite nach einen bestimmten string abfragen.

Also z.B. http://host01/test/ gibt als Antwort {"success":"true"}

Wenn ich diese String nun in vars.http_string = "{"success":"true"}" packe bekomm ich einen Fehler da hier " und { } im string vorkommen.

Kann ich das irgendwie umgehen?

LG mandelbrot

Icingaweb2 falsche Url beim Aufruf von Graphite

$
0
0
Hallo zusammen,

ich habe ein kleines Problem.
Ich habe Graphite installiert und wollte das Graphite Modul für Icingaweb2 benutzen.
Graphite an sich funktioniert einwandfrei, nur das Modul will nicht so recht.
Wenn ich bei einem Service (nehmen wir zum Beispiel ping4) auf das "rta" oder "pl" klicke, versucht mir das Modul folgende URL aufzurufen.
http://servername/graphite/render?&targe…ted&from=-10min
Als Ergebnis erhalte ich allerdings nur "No Data".
Nach ein paar "Nachforschungen" bin ich darauf gestoßen, dass in der URL ein paar Sachen fehlen.
Sie sollte eigentlich so lauten:
http://servername/graphite/render?target…fdata.rta.value ...

Liegt das an einer falschen Konfiguration meinerseits?

P.S
Icinga-Version: 2.0.0
Die Baumstruktur im Graphite sieht bei mir so aus :

[attach]7468[/attach]

im Verzeichnis icinga2:

[attach]7467[/attach]

Grüße

is it possible to process logs under /var/log/nagios/archives/

$
0
0
Great software. I really like the user interface. Installation was pretty easy after figuring out this little snag regarding cookies.

We chose the bulk mode and it seems to be working well, but only shows graphs from the date PNP4Nagios was live. We have four years of archives in Nagios. Is it possible to have pnp import those data for graphing and reports?

SNMP: Zusätzliche MIBs

$
0
0
Hi,

da ich mich gerade aufgrund des Updates auf die V1.3 etwas mehr mit OMD beschäftige: Wie erweitere ich die MIB Datenbank in OMD. Ist es möglich die MIB.txt (Bspw Sophos/Astaro) einfach in OMD abzulegen, damit diese richtig gedeutet wird oder wo setzt man dabei am besten an?

LG,

Chris

Anhang Astaro MIB

Nagvis gadget "pnp_graph.php" showing in tutorial not found in drop down gadget_url

$
0
0
Dear Expert

ref one tutorial , how to add graph in Nagios/Nagvis, I tried in our Nagios/Nagvis but there is no gadget called pnp_graph.php , no option available in the drop down. Appreciate attaching here.

https://www.youtube.com/watch?v=f66USWkNRJk

2014R2.0 / Nagvis 1.7

Regards

Graphite unter icigna 1 nutzen

$
0
0
Hi Leute,

ich habe folgendes Problem:

ich soll von der Arbeit her das Graphite Plugins testen. Wir nutzen Icinga 1.13 und werden wahrscheinlich auch nicht so schnell auf icigna2 wechseln.
Da sich niemand von uns mit Graphite auskennt habe ich im I-net nach anleitungen gesucht, doch die Beziehen sich alle auf icigna2, oder sind nur allgemeine Anleitungen zu Grahipe und zeigen nicht wie man es in icigna integriert.

meine frage ist nun ob es überhaupt möglich ist Graphte unter Icinga 1 zu nutzen?

Über Ratschläge wäre ich sehr dankbar

EDIT:

Ich habe Graphite zu laufen gebracht, aber wie ich es mit Icinga verknüpfen kann weiß ich immer noch nicht

Icinga2 - monitoring servers in different domains

$
0
0
Hello,

We have a lot of server in different domains separated using firewall. We are trying to implement a icinga2 environment which monitors all the servers using remote agents.
As a starting point, I am creating a POC with just 3 servers; one is master and other two correspond to 2 different projects (they belong to different domains).

Could any of you please guide me on how to implement icinga2 in such a way that, icinga running under master node gets the information from 2 remote nodes and this information is captured in icingaweb2 running in master node?

Do I have to install icinga2 in all the 3 servers or I have to install only icinga2 client in remote nodes?

Thanks,
Vishnu

icngaweb2 and graphite - perfdata in wrong format

$
0
0
Hi,

I've configured my very well running icingaweb2 server running under CentOS 7 (icinga version 2.3.10) with graphite going through this documentation http://dokuwiki.nausch.org/doku.php/cent…icinga:graphite.

The perfdata is written to directory /var/lib/carbon/whisper/icinga/.... and also at the graphite dashboard I'm able to see the graphs. At the icingaweb2 web interface all graphs are shown with an empty "no data" graph. I found out that the perfdata writer doesn't write the expected format <prefix>.perfdata.<perfdata-label>.value. Instead the the perfdata is formatted <prefix>.perfdata.<perfdata-label> (without value).

This causes that the graph urls at the icingaweb2 interface are not shown correctly.

Googling for almost 1 day and I didn't found a solution ?(

So any help appreciated.

Best regards
Dirk

Verständnisproblem: Konstanten und ihre Verwendung als Macros

$
0
0
Hallo zusammen,

ich versuche die Anzeige bestimmter custom attributes auf der Host-Seite (custom variables) zu unterdrücken (User, Passworte). Nach dem wälzen der Doku und etwas guhgeln kann ich die Verständnis-Lücke nicht schließen.

Damit Passworte etc. bei den Custom variables nicht angezeigt werden, sollen diese in der constants.conf als Konstanten definiert werden (so habe ich es verstanden). Tue ich dies und weise diese Konstanten custom attributes zu, so werden diese zur Laufzeit aufgelöst und trotzdem angezeigt. Die conf sieht in etwa so aus:

constants.conf:
...
dbuser = dbabc
dbpw = pass123
...

host.conf:
....
vars.dbmoniuser = dbuser
vars.dbsidpw = dbpw

servcice.conf:
...
vars.dbuser = host.vars.dbmoniuser
vars.dbpw = host.vars.dbsidpw

xy-commands.conf:
....
"--user" = "$dbuser$"
"--pw" = "$dbpw$"

Die Doku beschreibt unter dem Punkt Runtime Macros:

Evaluation Order

When executing commands Icinga 2 checks the following objects in this order to look up macros and their respective values:

User object (only for notifications)
Service object
Host object
Command object
Global custom attributes in the Vars constant

Nun habe ich die Konsante als Makro in xy-commands.conf eingetragen, was zur Fehlermeldung "non-optional macro..." führte. Dies ist aber auch nicht das Ziel, hierfür einen festen Wert für einer Command-Definition zu hinterlegen, nur damit er bei den Custom variables der Wert nicht angezeigt wird(weil ist praxisfremd). Die Anforderung ist, dass über jeden Host unterschiedliche Werte mitgegeben werden können und diese bei custom variables nicht sichtbar sind.

Wo liegt die meine Lücke bzw. der Denkfehler oder das Missverständnis?

Cluster Probleme und Error im Webinterface

$
0
0
Hallo,

bei uns läuft ein icinga2 HA-Cluster mit zwei Nodes.
Diese sprechen auch soweit miteinander und teilen die Checks auf.
Einer ist immer der aktive IDO-Writer.

Allerdings zeigt sich folgendes Fehlerbild:
Wenn ein Server neu gestartet wird holt dieser sich keinen Dump aus der Datenbank und gleich diesen ab sondern führt dann einmal alle Checks durch.
Das zeigt auch das Debuglog. Zudem scheinen sich beim Standard Check "proc" dadruch die Prozessanzahl zu addieren bis der nächste geplante check durchgelaufen ist.



Im Interface vom icingaweb2 habe ich folgenden Error stehen sobald ich einen Host aufrufe:
15.10. 08:48

ERROR

Stacktrace: #0
/usr/share/php/Icinga/Web/Controller/Dispatcher.php(55):
Zend_Controller_Dispatcher_Standard->dispatch(Object(Icinga\Web\Request),
Object(Icinga\Web\Response))

#1 /usr/share/php/Zend/Controller/Front.php(954):
Icinga\Web\Controller\Dispatcher->dispatch(Object(Icinga\Web\Request),
Object(Icinga\Web\Response))

#2 /usr/share/php/Icinga/Application/Web.php(333):
Zend_Controller_Front->dispatch(Object(Icinga\Web\Request),
Object(Icinga\Web\Response))

#3 /usr/share/php/Icinga/Application/webrouter.php(109): Icinga\Application\Web->dispatch()

#4 /usr/share/icingaweb2/public/index.php(4): require_once('/usr/share/php/...')

#5 {main}





15.10. 08:48

ERROR

Zend_Controller_Dispatcher_Exception in
/usr/share/php/Zend/Controller/Dispatcher/Standard.php:248 with message:
Invalid controller specified (logos)

Weiß da schon jemand Bescheid was zu tun ist?

Mit freundlichen Grüßen
R.

[Solved] HW/SW inventory zeigt in der Web-Gui keine Ergebnisse an

$
0
0
Hallo zusammen,

da für mich die Funktion des Hard- und Softwareinventory von großem Interesse ist, habe ich auf einem Testsystem die aktuelle stable-Version (1.2.6p12) von check_mk raw aufgesetzt
Testweise habe ich einen Clienrechner (Win 8.1 x64) ins Monitoring aufgenommen

Der liefert auch brav die Werte aller erkannten services ab, so daß ich sie mir an der Web-Gui anzeigen lassen kann.
nachdem ich das HW/SW Inventory-plugin (mk_inventory.ps1) in den Plugin-Ordner des Clients kopiert habe,
zeigt die Ausgabe von

Quellcode

1
cmk --debug -vv -i <hostname>
keine Fehler

Die ersten Zeilen lassen mich vermuten, daß das Plugin wie erwartet ausgeführt wird

Quellcode

1
2
3
4
5
6
7
8
9
10
11
Check_mk version 1.2.6p12
Connecting via TCP to 192.168.111.50:6556.
Added persisted section win_disks.
Added persisted section win_exefiles.
Added persisted section win_cpuinfo.
Added persisted section win_video.
Added persisted section win_system.
Added persisted section win_os.
Added persisted section win_bios.
Added persisted section win_reg_uninstall.
Added persisted section win_wmi_software.


Im Wato werden allerdings nach Hinzufügen/Starten des Plugins für den Testclient keine zusätzlichen Services angezeigt; das HW/SW Inventory bleibt leer.

Momentan habe ich keine Idee, wo ich weitere Fehlersuche betreiben könnte und würde mich über einen Hinweis freuen.
In diesem Zusammenhang die Frage: Ist es richtig, daß die Date "logstate.txt" im Verzeichnis des Agents auf dem Client ohne Inhalt ist ?

Gruß

ranx
Viewing all 1338 articles
Browse latest View live