Results 1 to 6 of 6

Thread: Distributed Shinken with more pollers

  1. #1

    Distributed Shinken with more pollers

    Hi,

    I have Shinken on a monitoring server and I need to monitor two different private networks. I have installed one poller in each network and I have communication between the monitoring server and the pollers servers. Also added the pollers as pollers_tags into Shinken configuration. The problem is how do I configure the scheduler from the monitoring server in order to communicate with both pollers at the same time? Currently I have something like this, and the scheduler communicates correctly, but only with poller-2:
    define scheduler{
    scheduler_name scheduler-1 ; just the name
    address 10.1.0.2 ; ip or dns address of the daemon
    port 7768 ; tcp port of the daemon
    spare 0 ; (0 = not a spare, 1 = is spare)
    weight 1 ; (some schedulers can manage more hosts than others)
    timeout 3 ; 'ping' timeout
    data_timeout 120 ; 'data send' timeout
    max_check_attempts 3 ; if at least max_check_attempts ping failed, the node is DEAD
    check_interval 60 ; ping it every minute
    modules PickleRetention
    realm All

    define poller{
    poller_name poller-2
    address 10.1.0.1
    port 7771
    poller_tags ptest1
    }

    define poller{
    poller_name poller-3
    address 10.2.0.1
    port 7771
    poller_tags ptest2
    }
    If I change "address *" for the scheduler I got errors for both pollers because they do not initialize the connection with the scheduler.

  2. #2

    Re: Distributed Shinken with more pollers

    Forgot to mention that the Shinken monitoring server has 3 networks interfaces including 10.1.0.2 and 10.2.0.2 . With the last two it communicates with the two pollers from the private networks.

  3. #3
    Administrator
    Join Date
    Jun 2011
    Posts
    216

    Re: Distributed Shinken with more pollers

    Check the *.ini file of the scheduler and pollers. You should have a line with 0.0.0.0 to say that they are listenning to all network interfaces. Another thing to check : version of Shinken. They have to be the same for all deamons

  4. #4

    Re: Distributed Shinken with more pollers

    Been there done, done that All the ini files have the the default "host = 0.0.0.0", and all are version 1.0.1. Actually if I try "address *" in Scheduler object definition the arbiter doesn't start anymore so I guess "*" is not a good value for the "address" parameter. Here it is the arbiter error:
    2012-03-16 10:22:46,838 [1331893366] Error : in creation connection for scheduler-1 : unknown host

    2012-03-16 10:22:46,839 [1331893366] CRITICAL ERROR: I got an unrecoverable error. I have to exit

    2012-03-16 10:22:46,839 [1331893366] You can log a bug ticket at https://github.com/naparuba/shinken/issues/new to get help

    2012-03-16 10:22:46,840 [1331893366] Back trace of it: Traceback (most recent call last):
    File "/usr/local/lib/python2.7/dist-packages/shinken/daemons/arbiterdaemon.py", line 468, in main
    self.do_mainloop()
    File "/usr/local/lib/python2.7/dist-packages/shinken/daemon.py", line 219, in do_mainloop
    self.do_loop_turn()
    File "/usr/local/lib/python2.7/dist-packages/shinken/daemons/arbiterdaemon.py", line 507, in do_loop_turn
    self.run()
    File "/usr/local/lib/python2.7/dist-packages/shinken/daemons/arbiterdaemon.py", line 605, in run
    self.dispatcher.dispatch()
    File "/usr/local/lib/python2.7/dist-packages/shinken/dispatcher.py", line 386, in dispatch
    is_sent = sched.put_conf(conf_package)
    File "/usr/local/lib/python2.7/dist-packages/shinken/satellitelink.py", line 110, in put_conf
    pyro.set_timeout(self.con, self.data_timeout)
    File "/usr/local/lib/python2.7/dist-packages/shinken/pyro_wrapper.py", line 115, in set_timeout
    con._setTimeout(timeout)
    AttributeError: 'NoneType' object has no attribute '_setTimeout'

    Maybe it is not possible what I am trying, or....is it ?!

  5. #5
    Administrator
    Join Date
    Jun 2011
    Posts
    216

    Re: Distributed Shinken with more pollers

    Yes only IPv4 (Never tried v6 maybe) are allowed in the field address.

    I think i found the problem. You have :

    shinken server : 10.1.0.2, 10.2.0.2, XXXXXXX (last interface you mentioned)
    poller-2 : 10.1.0.1 poller-3 10.2.0.1

    poller2 knows shinken server with 10.1.0.2
    poller3 knows shinken server with 10.2.0.2

    In the config file you specified the address of the scheduler :10.1.0.2. This is logic that poller-3 is lost .
    I guess you can't get the same IP address for the shinken server.

    Try to do another scheduler with the 10.2.0.2 address. Then you create 2 realms, one scheduler and one poller together. If you don't create realms, you will have errors in your logs telling you that the second scheduler cannot connect to poller-2 and vice versa.

  6. #6

    Re: Distributed Shinken with more pollers

    Thanks ! I was using only one realm (default : All).

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •