Archived community.zenoss.org | full text search
Skip navigation
7366 Views 12 Replies Latest reply: Mar 16, 2012 1:50 PM by dpetzel RSS
zack3ry Rank: White Belt 11 posts since
Nov 30, 2010
Currently Being Moderated

Mar 23, 2011 3:27 AM

Distributed Monitoring for Zenoss Core 3.x

Hi all,

 

Is there any documents for Distributed monitoring for Zenoss Core 3.x ? What I have were for Zenoss 2.4.x and it corrupted my exisitng zenoss server.

 

Would appreciate all the help if any.

  • jmp242 ZenossMaster 4,060 posts since
    Mar 7, 2007
    Currently Being Moderated
    1. Mar 24, 2011 8:17 AM (in response to zack3ry)
    Re: Distributed Monitoring for Zenoss Core 3.x

    Did you try the Zenpack?

    docs/DOC-5861

     

     

    --

    James Pulver

    Information Technology Area Supervisor

    LEPP Computer Group

    Cornell University

  • jmp242 ZenossMaster 4,060 posts since
    Mar 7, 2007
    Currently Being Moderated
    4. Mar 25, 2011 8:31 AM (in response to zack3ry)
    Re: Distributed Monitoring for Zenoss Core 3.x

    I'm sorry, I don't use distributed monitoring so I can't really be more help than to point you to the zenpack that says it works with v3. Maybe Egor will comment.

     

    --

    James Pulver

    Information Technology Area Supervisor

    LEPP Computer Group

    Cornell University

  • dpetzel Rank: Brown Belt 1,141 posts since
    Oct 17, 2010
    Currently Being Moderated
    5. Mar 25, 2011 9:14 AM (in response to zack3ry)
    Re: Distributed Monitoring for Zenoss Core 3.x

    I'm not sure about your specific error, but I did wade through setting up distributed collectors recently. I don't have a formal how to, but here are the high level steps from the notes I took. There maybe *better* ways to do it, but these were the steps I followed, and so far it appears to be working.

     

    1. Install Zenoss on your 'Main' Server' as if it were a standalone host. Ensure that box is working
    2. Install ZenPacks.community.DistributedCollectors on your main server
    3. Generate SSH Keys on your main server
      1. sudo su - zenoss
        mkdir ~/.ssh
        ssh-keygen -t rsa -f ~/.ssh/id_rsa -P ""
        chmod 700 ~/.ssh
        chmod go-rwx ~/.ssh/*
        cp ~/.ssh/id_rsa.pub /tmp/
        chmod a+r /tmp/id_rsa.pub

    4. Install Zenoss on your 'Remote Collector'. For now, pretend its another stand alone server, so install it just as you did in step 1 on the main server
    5. Open Firewall Rules between your main server and remote collector. docs/DOC-10246
    6. Add the Public Key from the keys you generated on the master server to the authorized_keys for the zenoss user on the remote collector.
    7. Test that SSH Key authentication is working correctly by logging into the main server as the zenoss user and attempt to ssh into the remote collector, you should be able to connect without needing to specify a password
    8. Start Zenoss (If you have not already). Let it run through its standard init process.
    9. Add the remote collector to Zenoss
      1. Log into the master server UI
      2. Navigate to Advanced --> Collectors
      3. From the gear menu, select "Add Remote Monitor"
      4. Enter the FQDN of the Remote Collector, and click OK (There might be a pause of several seconds)
      5. Place a check box next to the collector you just added
      6. From the gear menu, select "Update Remote Monitors
      7. You should get a message in the yellow bar indicating it was successfully updated.
    10. Restart Zenoss on the remote collector
    11. Disable MySQL on the remote collector. Because of the RPM dependencies, I couldn't remove MySQL.  So just turning it off

     

     

    In a nutshell, those were the steps I took.

  • felipe.castelar Rank: White Belt 14 posts since
    Jun 14, 2011
    Currently Being Moderated
    6. Jun 14, 2011 4:36 PM (in response to dpetzel)
    Re: Distributed Monitoring for Zenoss Core 3.x

    Thanks a lot for you post, dpetzel!

    After folowing this steps I get it working!

    One question: After restarting zenoss on remote collector, I can't access it from the web browser no more. Is this correct?

  • dpetzel Rank: Brown Belt 1,141 posts since
    Oct 17, 2010
    Currently Being Moderated
    7. Jun 14, 2011 10:29 PM (in response to felipe.castelar)
    Re: Distributed Monitoring for Zenoss Core 3.x

    That is correct, after the restart a few daemons are disabled, and your 'master' server becomes the only UI presence. It will communicate with the remote collectors to fetch the data it needs

  • m holler Newbie 3 posts since
    Feb 21, 2012
    Currently Being Moderated
    8. Mar 16, 2012 10:54 AM (in response to dpetzel)
    Re: Distributed Monitoring for Zenoss Core 3.x

    So I hate to revive an old thread, but I tried the method described here several different times, at first following the instructions verbatim (twice), and in subsequent tries I attempted to troubleshoot myself. I suppose the first question is whether the method dpetzel described still works on the current version (v.3.2.1)? If so, I have tried multiple times following these instructions to the proverbial "T" and still no dice. After finishing the install, I'd get to a point where I'd have a Zenoss "master" box with a web interface, and a collector without one. SSH between the machines was working correctly (a `ps aux` on the system would show the master box SSH'd into the collector successfully). I added a remote monitor in on the Master's web interface, and attempted to add a device to be monitored by the remote collector (it showed up in the "monitor" drop-down in the add device menu) but the job to create it on the remote collector was stalled. I had tracked down the process on the collector server to "zendisc" and its creation process. I couldn't find anything wrong with the logs. Just to be thorough, I took down the firewall on both systems, and SELinux is (and has been since I installed the system) off.

     

    My operating system is CentOS x86_64 and I installed Zenoss via RPM. Any ideas?

  • dpetzel Rank: Brown Belt 1,141 posts since
    Oct 17, 2010
    Currently Being Moderated
    9. Mar 16, 2012 11:37 AM (in response to m holler)
    Re: Distributed Monitoring for Zenoss Core 3.x

    The sterps were written against 3.0, but quick review seems they are still revelant.

     

    run a zenoss status on the master. At first read it sounds like zenjobs might not be running?

  • m holler Newbie 3 posts since
    Feb 21, 2012
    Currently Being Moderated
    10. Mar 16, 2012 12:07 PM (in response to dpetzel)
    Re: Distributed Monitoring for Zenoss Core 3.x

    I'm in the process of reinstalling the Zenoss Master, it turns out I ended up breaking it out of desparation the last time I tried working with this. It's been a little bit. I'll respond definitively in a little bit, but if I recall correctly, Zenjobs didn't start by default, and I do remember starting it manually before. I'm not sure this is the problem though, since the collector box was getting an initalization command from the master through SSH. Or am I misunderstanding something?

  • m holler Newbie 3 posts since
    Feb 21, 2012
    Currently Being Moderated
    11. Mar 16, 2012 1:30 PM (in response to m holler)
    Re: Distributed Monitoring for Zenoss Core 3.x

    zenerror.png

    Alright, we're back in action. The above error happens when I'm trying to create a device to be monitored by the collector. The job is always active (the green spinny thing is next to it in the "jobs" section until I force kill it), but nothing shows up in the output. I was only able to get the above output because I noticed that navigating away from the page flashes a quick (i.e., a few tenths of a second) messages on the screen before it loads the new page. When you go back to it, the phantom output is gone.

     

     

    As for zenoss status, here's the output from the master:

     

    [zenoss@localhost ~]$ zenoss status

    Daemon: zeoctl program running; pid=8287

    Daemon: zopectl program running; pid=8292

    Daemon: zenhub program running; pid=8332

    Daemon: zenjobs program running; pid=8369

    Daemon: zenping program running; pid=8431

    Daemon: zensyslog program running; pid=8470

    Daemon: zenstatus program running; pid=8476

    Daemon: zenactions program running; pid=8506

    Daemon: zentrap program running; pid=8604

    Daemon: zenmodeler program running; pid=8609

    Daemon: zenrender program running; pid=8644

    Daemon: zenperfsnmp program running; pid=8668

    Daemon: zencommand program running; pid=8702

    Daemon: zenprocess program running; pid=8729

    Daemon: zenwin program running; pid=8766

    Daemon: zeneventlog program running; pid=8804

    Daemon: zenjmx program running; pid=8833

     

    And here's the output from the collector:

     

    [zenoss@localhost ~]$ zenoss status

    Daemon: zenping program running; pid=4713

    Daemon: zensyslog program running; pid=4756

    Daemon: zenstatus program running; pid=4758

    Daemon: zentrap program running; pid=4796

    Daemon: zenmodeler program running; pid=4836

    Daemon: zenrender program running; pid=4862

    Daemon: zenperfsnmp program running; pid=4895

    Daemon: zencommand program running; pid=4921

    Daemon: zenprocess program running; pid=4947

    Daemon: zenwin program running; pid=4982

    Daemon: zeneventlog program running; pid=5018

    Daemon: zenjmx not running

     

    Note that I've had the problem with zenjmx before, and fixed it by installing java and adding hubhost and monitor argments to /opt/zenoss/etc/zenjmx.conf, but I wanted to show the output as a fresh install exactly the way you described.

  • dpetzel Rank: Brown Belt 1,141 posts since
    Oct 17, 2010
    Currently Being Moderated
    12. Mar 16, 2012 1:50 PM (in response to m holler)
    Re: Distributed Monitoring for Zenoss Core 3.x

    since zenhub is running on your master (per the output) this sounds like a connectivity issue:

    docs/DOC-10246

     

    Can you confirm 8789 is opened from your collector to your master (tcpdump verification would be great)

More Like This

  • Retrieving data ...

Legend

  • Correct Answers - 4 points
  • Helpful Answers - 2 points