Results 1 to 8 of 8

Thread: Status of SnmpBooster in 1.2.2

  1. #1

    Status of SnmpBooster in 1.2.2

    Hi there,

    I've been reading on the SnmpBooster documentation all over again as I plan to reinstall my Shinken environment in order to get it cleaner from previous attempts.

    How is the current status of SnmpBooster in 1.2.2? I see most of the documentation has not been updated since October 2012, sometimes early November 2012.

    I remember having many problems with memcached. Is this still a required component? Is there any trick to make it more reliable?

    Thanks in advance!

  2. #2

    Re: Status of SnmpBooster in 1.2.2

    Would love to see more information in regards to the integration between SnmpBooster and Graphite too...

  3. #3
    Administrator
    Join Date
    Dec 2011
    Posts
    278

    Re: Status of SnmpBooster in 1.2.2

    Hello,

    To my knowledge SnmpBooster is stable and functional with Shinken.

    There is not much to add in the documentation, as everything is there. I could add more examples, but the existing INI files already contain a lot of real work templates.

    If there are sections you would like to see improved or if you have questions, feel free to ask them.

    No further development work has been done at this point. You can see the status of SnmpBooster in the specification section of the wiki (SnmpBooster pages). It may become necessary to add Snmp v2c Get/Get-Next, but that would be in the course of this year.

    I will be adding some more devices to test with SnmpBooster, primarly a variety of Cisco devices, so that portion of genDevConfig which creates the configurations for SnmpBooster will improve a bit.

    These days I am busy on deploying and writing rules for HIDS.

    Concerning SnmpBooster and Graphite: SnmpBooster is an integrated Shinken acquisition module. Shinken can natively forward data to Graphite using the Graphite broker module. SnmpBooster has nothing to do directly with Graphite. Does that answer your question?

    Hope this gives you an update.

    xkilian

  4. #4

    Re: Status of SnmpBooster in 1.2.2

    Thanks for the update and for your answer in SnmpBooster+Graphite.

    Is anyone having problems with memcached on CentOS 6.3?
    To me it looks like dropping its information every now and then, maybe I need more tuning on how many objects it can handle in memory?

    I was using SnmpBooster running in a remote poller, maybe the documentation could be a little clearer on this point?
    This is the case where it seemed the most intermittent to me.

    Thanks

  5. #5
    Administrator
    Join Date
    Dec 2011
    Posts
    278

    Re: Status of SnmpBooster in 1.2.2

    Hmmm..

    Running a poller with a distant (not on the same LAN) memcached server is not a good idea. This is one area that we have not really stress tested. memcached can drop data.

    Probably need to add a monitoring piece to get stats on memcache to make sure we understand what is going on.

    genDevConfig just got some new patches to templates and a small bug fix.

    Any other feedback, patches, suggestions and corrections are welcome.

    Cheers,

    xkilian

  6. #6
    Junior Member
    Join Date
    Nov 2012
    Posts
    22

    Re: Status of SnmpBooster in 1.2.2

    [quote author=xkilian link=topic=747.msg4223#msg4223 date=1360031661]
    Hmmm..

    Running a poller with a distant (not on the same LAN) memcached server is not a good idea. This is one area that we have not really stress tested. memcached can drop data.

    [/quote]

    and memcached will drop data (cf memcached wiki : "Forgetting Data is a Feature&quot

    memcached must not be used to store data, but to cache data always available by other means (database, file, etc.)

  7. #7

    Re: Status of SnmpBooster in 1.2.2

    It would be great if we could get SnmpBooster work in distant remote pollers.

    Due to geographical constrains I'll be facing a 80-100 ms latency to some of my realms.

    At first thought I imagine running memcached on the distant pollers, as much of the communication will be between poller itself and its local memcached.
    Don't know if it's feasible, tho...


  8. #8
    Administrator
    Join Date
    Dec 2011
    Posts
    278
    This is still not in the plans. The key value store (no longer memcached but redis) must be network local to your pollers, latency would be bad in this case.

    I am sure it is feasible to implement, but no one is available to do it. For our uses, snmpbooster is doing the job to supervise lots and lots and lots of devices and OIDs. :_)

    openglx, if ever you see this post are you still using snmpbooster?!

Posting Permissions

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