• Performance Issue - Virtualized 2.3 under KVM

    5
    0 Votes
    5 Posts
    4k Views
    F
    @randyruiz: To add a further data point. I have a centos 7 vm spun up on the same host using virtio drivers going out of the same MACVTAP interface. This centos vm is giving me 970Mb as measured by IPERF in throughput. So this points further to freebsd and pfsense as where the problem lies. you might try changing rx offloading off as well (so rx off) and see if that gets the rest back, although usually only tx off is required. But yeah Freebsd is one of the outliers right now that drops these unchecksummed packets, most other OS'ses handle them without issue Of course make sure you're doing the tx/rx off for both the pfsense wan and lan virtio adapters not just one. If that's still not doing it I'd monitor cpu usage inside and outside of the VM and see if you're not hitting a bottleneck elsewhere Sorry I can't be of more help as I've never used KVM before. You might get more KVM specific help if you ask in the virtualization subforum
  • 2.3-RELEASE -> Status -> Traffic Graph numbers don't add up…

    3
    0 Votes
    3 Posts
    974 Views
    C
    The rate output is, and always has been, a little odd in that regard. It doesn't necessarily show everything depending, and doesn't show IPv6 at all. It's good for seeing when a specific IP is sending a lot of data. Beyond that, it's never been useful for much. The actual graph data is correct and includes IPv4 and IPv6, that pulls from the NIC's counters.
  • HTTP_ReFERER

    2
    0 Votes
    2 Posts
    812 Views
    jimpJ
    Install a browser extension like RefControl that lets you disable the HTTP_REFERER header on the browser, then login by IP address as usual.
  • HELP - Upgrading has Crippled Our VPN…

    3
    0 Votes
    3 Posts
    980 Views
    D
    @brandonpoc: I'd connect in and reinstall 2.2.6 on the virtual machine but I can't because I can't VPN in. I don't understand why it was necessary to completely nuke it (and not provide a package, as far as I can tell, for reinstalling it for those dependent on it). There is always a balancing act between convenience and security. It is not surprising that the pfSense team, as responsible vendors of security software, have taken the decision to remove older security standards that are known to be weak or compromised, also to configure system components to require security best practice. The removal of the PPTP server had been advertised for some time - there were numerous posts in the forums and it was clearly mentioned in the 2.3 release notes. There is an argument that the auto-updater ought to point the user to the release notes before allowing the install of a major upgrade, which has been suggested elsewhere in the forum. The new modularised structure of pfSense should allow for more frequent, smaller updates, which will help prevent the install shock of larger updates. Whether an update is large or small, I would argue that it is unwise to install a new version on a production firewall remotely without any form of physical or remote 'lights out' access. Upgrading one member of a clustered installation is not entirely foolproof, as it might prove impossible to force a failover to a working server remotely if the upgrade goes wrong. I believe it was a deliberate decision not to create a PPTP server package. Those users that had ignored the warnings in recent versions of pfSense to discontinue use of PPTP as a VPN protocol would likely have installed the package and continued to use PPTP, even if it was inappropriate for them to do so. Not every user is competent to evaluate whether PPTP is still appropriate for their environment in the light of the known brokenness. As has been said, you can either rely entirely on encrypted protocols (in which case a switch to SNMPv3 would be wise, if possible), or move to a supported VPN standard. divsys has given you a couple of suggestions for OpenSSL support in OS X. You might also like to consider IKEv2 - with carefully chosen parameters, it can work using the built in clients on a wide variety of OSes (though you may well find that the strongSwan app works better than the built in IKEv2 functionality on Android - you can find the app in the Google Play store). IKEv2 tends to be much less troublesome than IKEv1 IPsec. The historic security philosophy was to leave old, weaker standards enabled for compatibility. This approach became discredited and has been abandoned in recent years, as people were leaving weak and broken settings enabled, such as the pathetically weak 40 bit export cyphers in SSL, also the trivially broken WEP. These old standards provided no more than an illusion of security and created an unnecessarily broad attack surface. With many of the recent SSL/TLS attacks being downgrade attacks, the risk of leaving old standards in place implemented by poorly maintained code become clear: the time had come for these older standards to be retired and the code removed. The recent forced upgrade to SHA256 signatures in SSL/TLS certificates driven by the browser vendors is another example of forced abandonment of an older standard - in this case, the questionable SHA1 hash function. Other hardening in pfSense 2.3 can catch out the unwary. WEP has been removed from the wireless code, as it is utterly broken and there are few wireless devices still in use that do not support at least WPA (though it is best to use WPA2 with WPA mixed mode turned off if all devices support WPA2). TLS 1.0 is not supported any more due to security concerns, so those still following outdated advice to disable TLS 1.1 and TLS 1.2 in their browser will be unable to connect to the user interface in HTTPS mode. I got briefly caught out by the tighter requirements for key exchange imposed by the SSH2 server in 2.3 - I tried connecting to a newly upgraded box from an SSH profile on a secondary workstation that did not have modern DH and ECDH methods enabled, so I couldn't connect until I enabled a suitable key exchange method.
  • RESOLVED:e-mail notifications – it's not even trying

    4
    0 Votes
    4 Posts
    2k Views
    E
    Ok, this was resolved. I added several logging statements to the PGP files and found that pfsense thought it was in booting mode (it had been up for 14 hours or so already.) Rebooted and all is good.
  • 2.3-RELEASE -> Status -> Traffic Graph doesn't show ipv6 addresses…

    1
    0 Votes
    1 Posts
    316 Views
    No one has replied
  • Advice for home use

    13
    0 Votes
    13 Posts
    2k Views
    PippinP
    Enabling TRIM worked. Very nice (:
  • 0 Votes
    3 Posts
    4k Views
    M
    Steve, your suggestion worked like a charm. I feel a bit stupid for not figuring this on my own. Anyway, maybe it is a bug after all. When user configures interface to get IP from DHCP, it shouldn't take any other settings into consideration, doesn't it?
  • Help with some cron-jobs (I think…)

    1
    0 Votes
    1 Posts
    409 Views
    No one has replied
  • Major issues with DHCP (no leases available) 2.2.6 Had to repost

    2
    0 Votes
    2 Posts
    886 Views
    C
    Huh, not sure what happened to your old post, but it got weird. What do you get for: ls -l /var/dhcpd/var/db/
  • LACP teaming fail with pfsense 2.3 and cisco switch

    1
    0 Votes
    1 Posts
    936 Views
    No one has replied
  • Install third NIC and I lose WAN connection

    11
    0 Votes
    11 Posts
    2k Views
    P
    The 4-port Intel NIC fixed the issue.
  • 2.3 firewall rule state entry bug

    5
    0 Votes
    5 Posts
    2k Views
    H
    @xbipin it was fixed by Jorge M. Oliveira & steve_b please confirm by gitsync or waiting for the next round of snapshots https://redmine.pfsense.org/issues/6175
  • Dnsleaktest dns resolver unbound ? result

    2
    0 Votes
    2 Posts
    771 Views
    jimpJ
    Yep, unbound with forwarding mode off will query the roots and other authoritative DNS servers directly. That's the expected result in that configuration.
  • What happened to rc.create_full_backup in 2.3?

    4
    0 Votes
    4 Posts
    2k Views
    jimpJ
    It didn't "work so well" though, you just didn't happen to have any issues with it yet. The main problem is that the backup does not yield a clean filesystem. It's a BWOS (big wad of stuff) that gets spammed over the top of what you have now. If a file doesn't exist in your backup, but was there before the backup was restored, it's still there. That can wreak havoc on several areas that rely on loading things dynamically based off directory contents, and have other negative consequences. We've seen it break PHP before, and packages, just to name two big ones. Reinstalling+Restoring is so fast (Especially if you use PFI or 'rescue config.xml' in the installer)  that taking the risk rolling the dice with the old "full backup" just isn't worth it anymore.
  • Block proxy bypass

    2
    0 Votes
    2 Posts
    3k Views
    N
    Hi, I think what you did is useless for what you want to do but perhaps I misunderstand something you wrote. So Are you using transparent or non-transparent proxy? And are you using http and https on your proxy or only http? In general if this is a non-tranparent proxy configuration with http and https you have to configure something like this: 1.) From LAN to pfsense interface port 3128 (squid Port, will handle http and https traffic) 2.) From LAN to pfsense interface DNS (don't allow it to the internert. pfsense/squid will do the DNS lookup. Your client's browsers will just ask the squid proxy and it will do the rest) 3.) Block anything else from LAN to Internet but at least block http and https to ANY (except for your admin clients the need to have access to pfsense WebUI or you enable the "anti-Lockout" rule. In your browsers (Firefox, IE, Chrome) you have to enter the IP address of pfsense LAN interface and port 3128 for http, https and so on. This would be the way to avoid proxy bypass. Regards
  • WEB configuration port

    1
    0 Votes
    1 Posts
    457 Views
    No one has replied
  • Firefox refuses cert after 2.3 upgrade

    2
    0 Votes
    2 Posts
    1k Views
    jimpJ
    Check about:config, look at the value for security.tls.version.max It should be unset (default '3'), someone else seeing a similar error had somehow managed to set it to 1, and we now disable TLS v1 for security reasons.
  • BGP - OSPF - VLAN's…and everything in between

    2
    0 Votes
    2 Posts
    907 Views
    G
    @fusionp: Hi all, I'm a WISP and my current setup is multi-wan, I NAT over 8 load balanced WAN connections, I have 4 VLANS currently configured on my network, one of these I will be using for Captive portal on pfsense (right now it connects to a downstream mikrotik which routes over to my pfsense). My future plan is to change over from multi WAN to a single fiber connection, at this point I would like to move over to BGP so that I can give each user their own public IP. The likelihood is that I will need a total of 1000 addresses, so I may twist my ISP's arm so that they can issue me a /20 or /21 depending. I've tried to investigate BGP on pfsense but I still have a few questions, if I receive a /21 range from my ISP, can I divide this into 2x /22's so that I can issue a range of addresses for a particular vlan such as my captive portal vlan? I've seen some tutorials where OSPF is needed? In my case I will only be routing out through my pfsense…will I still need to use OSPF for any purposes? Will captive portal work with BGP routed instead of NAT? Please excuse my questions if they don't make total sense as I have zero experience with BGP. Any help much appreciated. Unless I misunderstand what you are doing, you don't any routing protocol at all - let your ISP deal with that. Simply subnet whatever range your ISP gives you into what ever ranges suit your needs.
  • Home Net

    5
    0 Votes
    5 Posts
    1k Views
    bmeeksB
    @TecTI: Problem solved. By selecting the custom_home_net in the pass list drop-down selector on snort interface I could block internal alerts source IPs. Thanks for your help. Yes, this part is key (selecting the desired custom list on the INTERFACE SETTINGS tab).  Simply creating a list on the PASS LIST screen is not enough.  You must then tell Snort (or Suricata, if using that package) to use the new list. Bill
Copyright 2025 Rubicon Communications LLC (Netgate). All rights reserved.