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

    Pfsense ignores/blocks machines after high traffic

    Scheduled Pinned Locked Moved General pfSense Questions
    14 Posts 4 Posters 4.5k 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.
    • chpalmerC
      chpalmer
      last edited by

      What kind of hardware are you running on?

      Sounds like state exhaustion but cant know until more info.

      :)

      Triggering snowflakes one by one..
      Intel(R) Core(TM) i5-4590T CPU @ 2.00GHz on an M400 WG box.

      1 Reply Last reply Reply Quote 0
      • B
        bollerup
        last edited by

        A hp U1 server

        2gigs of memory
        Version 2.0.2-RELEASE (amd64)
        built on Fri Dec 7 22:39:16 EST 2012
        FreeBSD 8.1-RELEASE-p13
        Platform pfSense
        CPU Type Intel(R) Pentium(R) D CPU 2.80GHz

        What info can i get for you ?

        1 Reply Last reply Reply Quote 0
        • chpalmerC
          chpalmer
          last edited by

          On your dashboard look for "states" like the picture below.  In my picture Ive just rebooted. (still getting calls asking why the internet is down…  ;D)

          The number on the right of the slash is maximum your box is set for.  The number on the left is present.  What are those?

          states.JPG
          states.JPG_thumb

          Triggering snowflakes one by one..
          Intel(R) Core(TM) i5-4590T CPU @ 2.00GHz on an M400 WG box.

          1 Reply Last reply Reply Quote 0
          • B
            bollerup
            last edited by

            Sorry, got confused there :)

            Last config change Thu Apr 4 20:46:09 CEST 2013
            State table size
            Show states
            MBUF Usage 2724/25600

            1 Reply Last reply Reply Quote 0
            • chpalmerC
              chpalmer
              last edited by

              The one you show is MBUF. What does the "State Table Size" above it show?

              The first number should increase as traffic increases.  You might try increasing the maximum states on    "System-Advanced-Firewall/NAT  Firewall Maximum States"

              Try 1million and see if your problem ceases.

              Triggering snowflakes one by one..
              Intel(R) Core(TM) i5-4590T CPU @ 2.00GHz on an M400 WG box.

              1 Reply Last reply Reply Quote 0
              • B
                bollerup
                last edited by

                Increased and let it run over night - same problem.. it kills the connection to any server with high traffic for XX minuts… something like 20 mins...

                1 Reply Last reply Reply Quote 0
                • J
                  josekym
                  last edited by

                  Go to Diagnostics->Tables and check your "virusprot" table to see if the IP addresses of the affected hosts are listed.  Maybe the high-bandwidth sessions are triggering the lockout via a firewall rule.

                  1 Reply Last reply Reply Quote 0
                  • B
                    bollerup
                    last edited by

                    Allready checked that, table is empty.

                    I tried to upgrade to latest dev. version - same issue.. still have to reboot the machine to get the systems back on track..

                    1 Reply Last reply Reply Quote 0
                    • chpalmerC
                      chpalmer
                      last edited by

                      Apr  4 19:50:16 pfsense check_reload_status: Syncing firewall
                      Apr  4 19:50:16 pfsense check_reload_status: Reloading filter
                      Apr  4 19:50:28 pfsense apinger: Error while feeding rrdtool: Broken pipe
                      Apr  4 19:52:33 pfsense check_reload_status: Syncing firewall
                      Apr  4 19:52:33 pfsense check_reload_status: Reloading filter
                      Apr  4 19:55:01 pfsense php: : Creating rrd update script
                      Apr  4 19:55:28 pfsense apinger: Error while feeding rrdtool: Broken pipe
                      Apr  4 20:00:02 pfsense php: : Creating rrd update script
                      Apr  4 20:00:28 pfsense apinger: Error while feeding rrdtool: Broken pipe
                      Apr  4 20:05:01 pfsense php: : Creating rrd update script
                      Apr  4 20:05:28 pfsense apinger: Error while feeding rrdtool: Broken pipe
                      Apr  4 20:08:50 pfsense php: : Creating rrd update script
                      Apr  4 20:09:28 pfsense apinger: Error while feeding rrdtool: Broken pipe
                      Apr  4 20:10:01 pfsense php: : Creating rrd update script
                      Apr  4 20:10:28 pfsense apinger: rrdtool respawning too fast, waiting 300s.
                      Apr  4 20:15:02 pfsense php: : Creating rrd update script
                      Apr  4 20:15:28 pfsense apinger: Error while feeding rrdtool: Broken pipe

                      Just noticed you added this to your first post after my reply…

                      Can you provide a topography of your setup?  Looks like your WAN port goes down.

                      What kind of connection?

                      Who is the provider?

                      What are your provider advertised speeds?

                      Ect-

                      Triggering snowflakes one by one..
                      Intel(R) Core(TM) i5-4590T CPU @ 2.00GHz on an M400 WG box.

                      1 Reply Last reply Reply Quote 0
                      • B
                        bollerup
                        last edited by

                        Hi,

                        Actually.. WAN stays online since i dont get disconnected from that.

                        Its hosted at Indicate in sweden to a 1gbit fiber connection - our ClearOS gateway servers are stable without issue… ironically - we firsted tested a virtual pfsense in hyper-v based on Alex's work - that one would crash and restart on high load.

                        Our provider have 2 x 1gbit fiber and we are connected directly to that.

                        Its worth to notice that i use nginx (also tested with squid) as a reverse proxy - however - there is no diffrence on wether i just use nat or the reverse proxies - it feels like i have i have tried every solution i can come up with and the logs dosent give any clear indication on why it happends.. to me it seems clear that pfsense is blocking the machine, since eg. server1 can be accesible while 2+3 is not.. first the servers is cut of.. then the webconfig becomes inaccesible .. then the machines sooner or later come back.. only to die 10-15 mins. later

                        1 Reply Last reply Reply Quote 0
                        • C
                          cmb
                          last edited by

                          Is it hitting the states limit? That's the exact symptoms of maxing out your state table. You can see in the RRD graph, System tab, states.

                          1 Reply Last reply Reply Quote 0
                          • B
                            bollerup
                            last edited by

                            states was set to 1mill limit earlier..

                            1 Reply Last reply Reply Quote 0
                            • B
                              bollerup
                              last edited by

                              Issue resolved by reinstalling the server with 32bit version.. seems to me that 64bit version of pfsense have driver issues with some network cards… or something like that.

                              Thank you for all of your help.

                              Bit of a note, webconfig still dies once in a while.. a 5 or 10 min. cronjob to restart the webconfig fixes that.

                              */5 * * * * /scripts/fixwebconfig.sh >/dev/null

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