Netgate Discussion Forum
    • Categories
    • Recent
    • Tags
    • Popular
    • Users
    • Search
    • Register
    • Login

    Combination of Squid Stable and Dansguardian causes very high MBUF

    Scheduled Pinned Locked Moved pfSense Packages
    5 Posts 3 Posters 2.0k Views
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • K
      kejianshi
      last edited by

      Hi all.  I was having a problem with very high MBUF approaching the limit.  Example  24460/25600.  This was fairly new.  Never happened before.  Normally, it would sit around  3460/25600 or so for many many days and never really change much.  The only things I had changed is I added squid stable and dansguardian to an otherwise absolutely vanilla setup.  So, I wiped the box and re-installed.  With no packages loaded, there was no MBUF issue.  After adding squid stable, there was still no MBUF issue.  After then installing Dansguardian, the MBUF problem returned.  So, I removed Dansguardian again leaving only squid stable and the MBUF numbers are back where they have always been.  Low. So, I figure this must be an issue with Dansguardian causing some sort of memory leak.
      Anyway, it would be nice is it didn't do that because I like dansguardian. 
      I'm using Pfsense version:
      2.0.3-RELEASE (i386)
      built on Fri Apr 12 10:22:57 EDT 2013
      FreeBSD 8.1-RELEASE-p13

      1 Reply Last reply Reply Quote 0
      • jimpJ
        jimp Rebel Alliance Developer Netgate
        last edited by

        Just increase your nmbclusters so you get more. Many workloads will max out the default mbuf values.

        http://doc.pfsense.org/index.php/Tuning_and_Troubleshooting_Network_Cards

        Remember: Upvote with the 👍 button for any user/post you find to be helpful, informative, or deserving of recognition!

        Need help fast? Netgate Global Support!

        Do not Chat/PM for help!

        1 Reply Last reply Reply Quote 0
        • K
          kejianshi
          last edited by

          OK - I'll do I'll try that tweak.  System has 1 nfe0 that I'm using on WAN that is 10/100.  The internet never gets to 100 much here.  I have 1 dual port PCIe Intel NIC that is 10/100/1000 and 1 single port PCI Intel NIC that is 10/100/1000.  Those are em0-em2.  The LANs are all disallowed to see each other to give my tenants privacy.  All The gigabit Intel NICs are set up as LAN ports.  So, will I have to apply this tweak for just the nfe(0) wan, the em(0-3) LANs or for all? I noticed the patches are different for each card type.

          1 Reply Last reply Reply Quote 0
          • A
            adam65535
            last edited by

            Does anyone know if the patch mentioned in the thread below will be integrated with pfsense at some point (either 2.1.x or 2.0.x)… or is it already?

            http://comments.gmane.org/gmane.os.freebsd.devel.net/39495

            It sounds like they fixed at least the igb driver to scale hw.igb.num_queues (and maybe mbufs?) better by default according to that somewhat recent thread (june 20th 2013).  I don't know if other drivers got the same fix or not.  I use quad port nics btw.

            1 Reply Last reply Reply Quote 0
            • K
              kejianshi
              last edited by

              A can confirm that making the changes helped MBUF.  No more issues.  I didn't have any erratic behaviour so I didn't apply the NIC specific fixes.  Just the increased MBUF.

              1 Reply Last reply Reply Quote 0
              • First post
                Last post
              Copyright 2025 Rubicon Communications LLC (Netgate). All rights reserved.