How to optimize NMIS

 
1
0
-1

My Metrics graph does not update and the Networks Metrics graph is empty. I have around 2600 routers and switches with around 13000 ports for collection (ports without # in description)

My NMIS server has 56 cores and over 250GB memory and cpu load average is less than 2 percent. I have set up every 5min ping and SNMP collection every 15min.

I dont get why there are so many overdue jobs. I have installed NMIS9 with default options.

Can I do anything to increase the amount of active jobs and get the Metrics graph to start updating?

    CommentAdd your comment...

    1 answer

    1.  
      2
      1
      0

      Hi Fredrik, 

      It seems that the NMIS workers should be increased. You can do this in the nmis9 Configuration file, in nmis9/conf/Config.nmis with the following parameter: 


      'nmisd_max_workers' => 10,


      After the change, you should restart nmis9 daemon, nmis9d. 

      Regards,

      1. Fredrik Karlsson

        So I went from 10 to 100 and now I start wondering how many workers should I have? 4x the cores?

        I will let it stabilize for a few hours

      2. Fredrik Karlsson

        It did not help, but I'm removing all devices and re-adding them to clear all jobs.

      3. Fredrik Karlsson

        I increased the worker count to 100 and after I deleted all devices it worked. I had alot of jobs that took over 15 min to clear each. Thanks for the help!



      4. Josune Cordoba

        Thanks for let us know! 

        For more detailed options with schedule, you can have a look at NMIS 9 Administration Notes


        Regards, 

      CommentAdd your comment...