Posts by MarcusCaepio

This forum was archived to /woltlab and is now in read-only mode.

    sru , ja das funzt. Hab ich schon getestet.

    2 Server, beide mit separater DB, einer config master, anderer Slave, enable_ha aus, beide Icingaweb2.

    Da beide die Check results bekommen, egal wer gerade den check ausführt, schreiben die das ergebnis in Ihre DB.

    Was eben vorkommen kann: minimalste Abweichung der Timestamps.

    Was aber auf einem ganz anderen Blatt steht, ist die Config Synchro der beiden Icingaweb2 Instanzen.

    Da gibt es nämlich nix und man muss sich selber was basteln oder immer beide Server anpacken.

    Wenn man immer genau einen Server hat, auf dem Icingaweb2 samt DB läuft, aber eine riesige Umgebung überwacht, läuft man in Gefahr, dass dieser Server irgendwann in die Knie geht? Immerhin schaufeln ja im schlimmsten Fall XXX satelliten ihre Check Results über ihn in die DB oder hab ich hier grad nen Denkfehler?

    Being with you sru.

    I also would call this a "bug" or maybe better " unfortunately one more issue which pretends me of using a icinga2 cluster". depending on the "Server X executes the check itself but X&Y&Z get the result" it surely is the common behaviour....unfortunately.....

    Imho there are several ways to solve this (by the Devs):

    - Only the Checksource server is allowed to trigger the eventhandler. We already get the Checksource per service, when we run a cluster, so what about a dependency between this and the event?

    - As we are going to build cluster in Top-Down Config. where we have always a config master, only the configmaster is allowed to execute events. But I don't know what happens, when you configure a second master for failover or sth. else (like 6.10.2. in icinga2 doc).

    - the dirty way: only one server has the event handler scripts stored? (what happens at the other two?)

    - just a thought in my head (which also could be implemented for standalone icingas) could be a more intelligent eventhandler. Instead of just being "stupid executing script xyz" what about adding a kind of syntax and response listener to the eventhandler itself? So you could check, if the execution of the script was successfull, depending on the exit code ( and the quality of the script of course) and you can build some kind of order: server 1 runs script, if fail server 2 runs script, if fail -> notifiaction/alert. This would also be another more reason to display configured eventhandler in Icingaweb2 (maybe with timestamp of last run) as since my last use of Icingaweb2 there was no display of eventhandlers although they are stored in the DB already.

    Anyway, thanks for the information jttrzebiato . This is a serious problem in my view and should really get attention. Automatic handling of problems by event handlers let's many of us sleep at night. The triggering of an eventhandler multiple times because of one error is a no-go.

    Hi All,

    I want to receive traps via logstash's snmptrap plugin, which itself works. Only problem is the missing MIBs.

    So I started to search how to add Vendor MIBs into the ruby-snmp environment, which is used by the snmptrap plugin and found this URL:

    Unfortunately, the import script doesn't seem to work.

    Doesn't matter which MIB I am trying to import, it fails with the message:

    smidump: module `mibs/XYZ.mib' contains errors, expect flawed output

    All required packages are installed. Logstash is running in version 5.4 on Ubuntu 16.04. libsmi in 0.4.8.

    So is anybody of you using this plugin and already imported Vendor MIBs?

    I kindly ask not to answer with sth. like "use filebeat+snmptt+grok" or sth. like this. I am also watching in this direction, just checking several options to figure out which matches my needs best.



    you could always just set the messurement on the axis to the right one. Grafana will then do the conversion

    Nope you can't because the graphs are generated depending on the host. When Server 1 has disk as first check = first graph and you set the unit to Bytes, change the server and server 2 has e.g. Cluster as first check = first graph, also the unit is Byte instead of linear or sth. else. It seems not assigned to the service name, but to graph's position.

    Hi all,

    Icingaweb 2.4.1

    it seems, that since my update to 2.4.1, there are files missing in the dompdf vendor folder. When I try to export to pdf, I get:

    1. Warning: Uncaught exception 'ErrorException' with message 'require_once(/usr/share/icingaweb2/library/vendor/dompdf/include/ failed to open stream: No such file or directory' in /usr/share/icingaweb2/library/vendor/dompdf/ Stack trace: #0 /usr/share/icingaweb2/library/vendor/dompdf/ Icinga\Application\ApplicationBootstrap->Icinga\Application\{closure}(2, 'require_once(/u...', '/usr/share/icin...', 52, Array) #1 /usr/share/icingaweb2/library/vendor/dompdf/ require_once() #2 /usr/share/php/Icinga/File/Pdf.php(15): require_once('/usr/share/icin...') #3 /usr/share/php/Icinga/Application/ClassLoader.php(306): require('/usr/share/php/...') #4 [internal function]: Icinga\Application\ClassLoader->loadClass('Icinga\File\Pdf') #5 /usr/share/php/Icinga/Web/Controller/ActionController.php(536): spl_autoload_call('Icinga\File\Pdf') #6 /usr/share/php/Icinga/Web/Controller/ActionController.php(481): Icinga\Web\Controller\ActionController-> in /usr/share/icingaweb2/library/vendor/dompdf/ on line 52
    2. Fatal error: main(): Failed opening required '/usr/share/icingaweb2/library/vendor/dompdf/include/' (include_path='/usr/share/icingaweb2/library/vendor:.:/usr/share/php:/usr/share/pear') in /usr/share/icingaweb2/library/vendor/dompdf/ on line 52

    In fact, the file /usr/share/icingaweb2/library/vendor/dompdf/ is missing.

    1. /usr/share/icingaweb2/library/vendor/dompdf: ls -l
    2. total 28
    3. -rw-r--r-- 1 root root 767 Jan 18 15:27
    4. drwxr-xr-x 7 root root 4096 Mar 9 09:37 lib
    5. drwxr-xr-x 11 root root 4096 Mar 9 09:37 src

    But why? I did not delete anything :D

    I know, PDF export worked fine in former versions. Any ideas?



    Könntest auch mit 3 weiter machen, aber dann eben mit 2 zonen.

    Ich habe in der master zone meinen configmaster+db+icingaweb+grafana

    und eine Zone "darunter" 2 Satelliten. Die Satelliten bekommen ihre Config vom Master, führen die eigentlichen Checks loadbalanced aus und alles wandert nach oben in den Master. Dann kannst du alternativ sogar noch einen 2. Configmaster daneben setzen. Nur wird es dann Tricky mit dem Syncen im Icingaweb....

    Instead of creating an action url in Icinga2, you can create a Navigation URL in Icingaweb2 and share this to all users.

    In your admin account go to:

    Your Name -> My Account -> Navigation -> Create new Navigation Item:

    Name: Grafana

    Shared: yes

    Type: Host Action

    Target: New Window

    Url; your grafana Url

    Icon: chart-area

    Same for Service Action



    Ich frage mich gerade wie? Also wie hast Du was gemacht, damit ein Eintrag hier für alle User automatisch vorhanden ist?

    Danke Dir im Voraus.


    Steht doch da ^^

    In /etc/icingaweb2/enabledModules/monitoring/config.php ganz am ende sind die default dashboards definiert. Die habe ich mir einfach nach meinen Bedürfnissen angepasst. Nur vorher bedenken, dass sie bei nem update wieder überschrieben werden, also vorher sichern ;)