Cacti interface bandwidth monitoring, incorrect value…



  • Hi all,

    I'm expecting some troubles to make proper cacti graph for each pfsense interface (vlan in my case). My Pfsense is a virtual machine.

    I enabled SNMP in Pfsense, and I create a generic snmp device in cacti. I make a verbose query. Looks good.

    Then I create a graph for each interface, works great.

    But when making comparaison with RRD graph in pfsense, I found that all my outbound value in graph (inbound seems ok) are twice as real value.

    I have no idea why. I try to make 64 bits counter but still the same problem.

    Thanks in advance for any clue.

    The correct graph taken from a Cisco 3560 Switch correlated by the RRD Pfsense graph

    The incorrect graph taken from pfsense with cacti



  • I am seeing this as well, and I can't seem to find that anyone figured out a solution. (using MRTG, not cacti, but whatever).
    Oddly enough, it is only happening on one of our pfSense installs (we have 11 running 2.2.1).  Of course, it is the one at our main corporate office, which means I'd like to get meaningful non-doubled bandwidth usage numbers from SNMP.

    One thought I had - this is also the only box we have running ipv6 and ipv4.  All of the others are ipv4 only.
    It is also the only one that is a OpenVPN server (the others are all clients).  All interfaces are done as VLANs (but I have that in 3 other offices with no problems).

    Any thoughts?  Are you running OpenVPN or IPv6?



  • This may be a bug in BSD.

    The RRD graphs had a similar problem where the WAN graph would show "real" traffic metrics, but the LAN graph would show 2x WAN.  I think the root cause of that was a bug in the version of BSD that pfSense uses.

    The two may be related, but I am not entirely sure.  The RRD graphs in 2.2 report properly.



  • @tim.mcmanus:

    This may be a bug in BSD.

    The RRD graphs had a similar problem where the WAN graph would show "real" traffic metrics, but the LAN graph would show 2x WAN.  I think the root cause of that was a bug in the version of BSD that pfSense uses.

    The two may be related, but I am not entirely sure.  The RRD graphs in 2.2 report properly.

    Ok, maybe I can clarify this some.  I'm running 2.2.2 (now) and still have the problem.  Yes, the RRD graphs in pfSense are correct.  However, like many people, I use SNMP to monitor all of my routers (in this case, via MRTG).

    SNMP numbers are incorrect on the router at my main site, and now at one remote site.  I am using OpenVPN.  I can't just reverse LAN numbers, because I'm multi-WAN at most sites - I want to see separate bandwidth for each WAN connection.  When I say incorrect, in this case, it means that my outbound traffic numbers are doubled.  Inbound traffic is correct.

    So, the second site started showing this symptom after I started messing about with the traffic shaping that was on at the remote site.  Now, even with traffic shaping off, I am seeing the doubled numbers.  I can't fathom why messing about with traffic shaping would cause this to start…but it appears that it did.



  • @althornin:

    @tim.mcmanus:

    This may be a bug in BSD.

    The RRD graphs had a similar problem where the WAN graph would show "real" traffic metrics, but the LAN graph would show 2x WAN.  I think the root cause of that was a bug in the version of BSD that pfSense uses.

    The two may be related, but I am not entirely sure.  The RRD graphs in 2.2 report properly.

    Ok, maybe I can clarify this some.  I'm running 2.2.2 (now) and still have the problem.  Yes, the RRD graphs in pfSense are correct.  However, like many people, I use SNMP to monitor all of my routers (in this case, via MRTG).

    SNMP numbers are incorrect on the router at my main site, and now at one remote site.  I am using OpenVPN.  I can't just reverse LAN numbers, because I'm multi-WAN at most sites - I want to see separate bandwidth for each WAN connection.  When I say incorrect, in this case, it means that my outbound traffic numbers are doubled.  Inbound traffic is correct.

    So, the second site started showing this symptom after I started messing about with the traffic shaping that was on at the remote site.  Now, even with traffic shaping off, I am seeing the doubled numbers.  I can't fathom why messing about with traffic shaping would cause this to start…but it appears that it did.

    Just an FYI - I can report that upgrading from 2.2.2 to 2.2.4 has resolved this for me on all of my pfSense routers.
    SNMP is now reporting correct values.


Log in to reply