Process named causing troubles

Discussion in 'Server Operation' started by kernelpanic, Jun 25, 2014.

  1. kernelpanic

    kernelpanic New Member

    hello guys, this is my first time here as a registered user.
    first of all, i wanna say sorry about my english, i talk spanish.
    i come with a big problem..
    the thing is:
    i buy a vps from volumedrive at http://volumedrive.com/vdrive/?a=vps is the package 2.
    i follow this tutorial http://www.howtoforge.com/perfect-server-debian-wheezy-apache2-bind-dovecot-ispconfig-3 and everything goes well.
    two days ago volumedrive suspend this vps because process named was causing high resource usage.
    after a long day they give me a new vps because the old one was "rare". none of us was able to make an os-reload in this vps didn't know why...
    anyway i ask volumedrive about resource limitations and they told me this:

    If the VPS causes such an issue that other VPS on the node are affected we cannot wait for a response from you. The load average limit is 5. Anything above that it gets suspended. You need to monitor your own VPS.

    two things: 1°, what means 5 and how to monitor this? and 2°, what can i do whit bind to decrease resource usage?

    i have read this articles and seems to work for these users:
    https://bugzilla.redhat.com/show_bug.cgi?id=709205#c29
    http://stackoverflow.com/questions/13059014/bind-named-service-high-cpu-load
    can you guys tell me some "solution" for this?
    if i use mydns instead bind i will se same error ¿?
    i was a w2k user sense 1° time, but now in linux im in troubles :)
    thanks in advance.

    pd: one of the first support ticket i send when vps was suspended they answered this:

    your VPS was suspended for causing the node to spike to a 17 load average.
    process causing it:

    [root@vps8 ~]# vzpid 623402
    Pid CTID Name
    623402 373 named
     
  2. till

    till Super Moderator Staff Member ISPConfig Developer

  3. kernelpanic

    kernelpanic New Member

    wow! it's the first time someone in a forum respond me that fast!
    thanks!
    i will update this post after i check your solution.
    in a few words it's recomended turn off recursion right?
     
  4. till

    till Super Moderator Staff Member ISPConfig Developer

    Yes. Otherwise your server might be misused for dns based dos attacks which might explain the high dns usage.
     
  5. kernelpanic

    kernelpanic New Member

    ok. i have made the customizations.
    how can i test now if this is working?
     
  6. till

    till Super Moderator Staff Member ISPConfig Developer

  7. kernelpanic

    kernelpanic New Member

    ok, if i put:
    allow-transfer {“none”;};
    allow-recursion {“none”;};
    recursion no;
    in resolv.conf.options bind doesn't start.
    and openresolver.com shows: Recursive resolver is not detected on xxx.xxx.xxx.xxx with or without modifications in resolv.conf.options
    so, in this case openresolver seems to not work.
    if i put just: recursion no; bind start ok.

    my scenario is:
    debian wheeze x64
    the perfect server with ispconfig3
    all apdated & upgraded.
    3 sites in the server working ok.
     

Share This Page