PfBlockerNG v2.0 w/DNSBL

tonymorella

OK here's a strange occurrence…
pfSense just upgraded to 2.3.2 but this behaviour happened with 2.2.1 on a physically different router.

pfb version 2.1.1_4.

On our LAN when we use google chrome (latest) and go to google it can take several seconds for google to display. Doing a search can take 20 seconds to respond, but it does come back eventually.

On firefox or any other browser, internet access is instant, including identical searches on the same PC simultaneously.

pfb is working OK when enabled ie blocking incoming and outgoing ipv4 correctly.

When pfBlockerNG is disabled, searches and access on chrome become instant too, so something in pfBlocker is slowing down just chrome.
This behaviour is consistent across different PCs on different versions of windows.

I don't really have much idea what this can be, and logs aren't telling me much.

tl;dr - chrome+pfblockerng = very slow. Firefox = fast. Turn off pfblocker =all browsers incl chrome fast

What block lists are you running? Also, ctrl-shift-I to get into dev mode for both look at performance in FF and timeline in chrome, run them both and compare.

robatwork

@tonymorella:

What block lists are you running? Also, ctrl-shift-I to get into dev mode for both look at performance in FF and timeline in chrome, run them both and compare.

Currently running top_v4 and top_v6 by selecting various countries in the Top 20.
Also 5 lists in IPv4 to block some social networking sites by AS number.
And DNSBL is enabled although I was testing the speed prior to this and I don't think it made any difference. TLD isn't ticked. All other options are default.

BBcan177

@robatwork:

On our LAN when we use google chrome (latest) and go to google it can take several seconds for google to display. Doing a search can take 20 seconds to respond, but it does come back eventually.

You can open the Chrome cache viewer with this URL (Clear that and see if it helps):

chrome://net-internals/#dns

Also ensure that the LAN devices only have pfSense as its DNS server.

robatwork

@BBcan177:

@robatwork:

On our LAN when we use google chrome (latest) and go to google it can take several seconds for google to display. Doing a search can take 20 seconds to respond, but it does come back eventually.

You can open the Chrome cache viewer with this URL (Clear that and see if it helps):
chrome://net-internals/#dns
Also ensure that the LAN devices only have pfSense as its DNS server.

Will try the cache clear, thanks.

However the LAN devices are all set with our domain servers as DNS servers. The DNS on the AD servers are all set to just have the pfsense as the only forwarder. pfSense has DNS Resolver only. I did just watch the last hangout about DNS and believe that I am setup OK with just resolver even though we are multi-WAN. If you think I'd be better off with DNS forwarder instead please let me know.

Thanks for your help :)

BBcan177

@robatwork,

So to be clear, ensure that the LAN devices have only your AD DNS servers defined. Then set the AD DNS server DNS Forwarder settings to pfSense only… This way DNSBL will filter the traffic.

The resolver can be in Resolver mode or in Forwarder mode.... I recommend Resolver mode....

If you are referring to the DNS Forwarder (dnsmasq), then DNSBL will not function, as its not configured for that.

robatwork

@BBcan177:

@robatwork,

So to be clear, ensure that the LAN devices have only your AD DNS servers defined. Then set the AD DNS server DNS Forwarder settings to pfSense only… This way DNSBL will filter the traffic.

The resolver can be in Resolver mode or in Forwarder mode.... I recommend Resolver mode....

That is exactly how we are configured (in Resolver mode). Testing continues….

jim82

Hi all,

Yesterday I added 4 new lists to DNSBL, see attachment 1.

Right after I added these new lists, my DNSBL Alerts/Log stopped working, and all blocks are now shown in the "DENY" log section. See attachment 2.

Can anyone shed light as to what I have done wrong?

Thanks

BR Jim

1.JPG_thumb

2.JPG_thumb

BBcan177

Hi jim82,

In the DNSBL tab, only add DNSBL based feeds, the RW_IPBL is an IP based list that should be added to the IPv4 Tab.

Goto the General Tab, and enable Suppression and follow that with a Force Reload - All… This will remove any RFC1918 or loopback addresses that might be in a list. I am going to make this standard in the next release to avoid this issue... The Deny alerts that you see are from the DNSBL_IP firewall rule. In DNSBL, you enabled the "DNSBL_IP" option which will collect and block any IP addresses that are found in a Domain based feed.

jim82

Thanks a lot for your swift reply! Should I remove the RW_IPBL list from DNSBL?

BR Jim
EDIT: I've now removed the list from DNSBL and added it to IPv4, is this the correct way of doing it?

@BBcan177:

Hi jim82,

In the DNSBL tab, only add DNSBL based feeds, the RW_IPBL is an IP based list that should be added to the IPv4 Tab.

Goto the General Tab, and enable Suppression and follow that with a Force Reload - All… This will remove any RFC1918 or loopback addresses that might be in a list. I am going to make this standard in the next release to avoid this issue... The Deny alerts that you see are from the DNSBL_IP firewall rule. In DNSBL, you enabled the "DNSBL_IP" option which will collect and block any IP addresses that are found in a Domain based feed.

Qinn

@RonpfS:

The tables are built from MaxMind GeoIPLite2 database, pfblockerNG just take the db and create the files for it's usage. MaxMind support has been contacted about the size being 3x larger than before.

Just out of curiosity, is there any progress on this? I saw that MaxMind: Last-Modified: Tue, 02 Aug 2016 is still on August. Of course I could do a "php /usr/local/www/pfblockerng/pfblockerng.php dc", but everything is working fine (pfBlockerNG v2.1.1_4) so no need for it, as in ….if it ain't broke, than don't try to fix it ;)

Cheers Qinn

varazir

Hello I keep getting this in the update log

Could not open ISO

UPDATE PROCESS START [ 10/09/16 10:30:37 ]

===[  DNSBL Process  ]================================================

[ hphost_partial ]	 exists.
[ mvps_hosts ]		 exists.
[ SomeoneWhoCares ]	 exists.
[ BBcan177 ]		 exists.
[ DNSBL_IP ]		 Updating aliastable... 
  no changes.
  Total IP count = 1

===[  Continent Process  ]============================================

Could not open ISO [ SH_v4 ]

[ pfB_Africa_v4 ]	 exists.
Could not open ISO [ AP_v4 ]

Could not open ISO [ CX_v4 ]

Could not open ISO [ CC_v4 ]

[ pfB_Asia_v4 ]		 exists. [ 10/09/16 10:30:38 ]
Could not open ISO [ PN_v4 ]

[ pfB_Oceania_v4 ]	 exists.
[ pfB_SAmerica_v4 ]	 exists.
[ pfB_Top_v4 ]		 exists.
[ pfB_Top_v6 ]		 exists. [ 10/09/16 10:30:42 ]
[ pfB_PS_v4 ]		 exists.

===[  IPv4 Process  ]=================================================

===[  IPv6 Process  ]=================================================

===[  Aliastables / Rules  ]==========================================

No changes to Firewall rules, skipping Filter Reload
No Changes to Aliases, Skipping pfctl Update

 UPDATE PROCESS ENDED [ 10/09/16 10:30:43 ]

BBcan177

@varazir:

Could not open ISO [ SH_v4 ]
Could not open ISO [ AP_v4 ]
Could not open ISO [ CX_v4 ]
Could not open ISO [ CC_v4 ]
Could not open ISO [ PN_v4 ]

Yes this is an issue with the MaxMind monthly database changes not reporting on some GeoIPs… I have a fix for this which will be in the next release which will add a "placeholder" for all GeoIPs regardless if they are not defined by MaxMind... Just ignore for the time being...

BBcan177

@jim82:

Thanks a lot for your swift reply! Should I remove the RW_IPBL list from DNSBL?
EDIT: I've now removed the list from DNSBL and added it to IPv4, is this the correct way of doing it?

Yes… remove from DNSBL and Add the RW_IPBL to the IPv4 tab...

BBcan177

@Qinn:

Just out of curiosity, is there any progress on this? I saw that MaxMind: Last-Modified: Tue, 02 Aug 2016 is still on August. Of course I could do a "php /usr/local/www/pfblockerng/pfblockerng.php dc", but everything is working fine (pfBlockerNG v2.1.1_4) so no need for it, as in ….if it ain't broke, than don't try to fix it ;)

Do you have any MaxMind update errors in /var/log/pfblockerng/geoip.log?

I would manually run the update and see if it reports any errors…

php /usr/local/www/pfblockerng/pfblockerng.php dc

regne

Hi.

I keep watching in the general system log entries like this:
*nginx: 2016/10/10 19:04:39 [error] 23499#100098: 737 open() "/usr/local/www/utsync.ashx" failed (2: No such file or directory),client: 10.10.10.1, server: , request: "GET /utsync.ashx?eid=50052&et=0&fp=2X9bJ6tnRz5B2L0llgZVTWayaMg4TcNYwOj-CDyEPl1k&return=http%3A%2F%2Fps.eyeota.net%2Fmatch%3Fbid%3Dr8hrb20%26uid%3Dnil HTTP/1.1", host: "ml314.com", referrer: "http://viraliq.com/15-musicians-didnt-know-passed?utm_source=revcontent&utm_medium=referral&utm_campaign=desktop&utm_content_id=996712&utm_boost_id=116872&utm_targeting=editorial%20news&utm_widget_id=31653"

I understand that DNSBL has successfully diverted the DNS petition to server 10.10.10.1 and the requested file is not there but is, logging these messages, the right behavior?

Is there a way to disable them?. They are populating the system logs and hiding important stuff.

erwintwr

just wanted to give a personal thank you to bbcan for his committed FREE support on this amazing addon.

you sir are a legend :)

Qinn

@BBcan177:

@Qinn:

Just out of curiosity, is there any progress on this? I saw that MaxMind: Last-Modified: Tue, 02 Aug 2016 is still on August. Of course I could do a "php /usr/local/www/pfblockerng/pfblockerng.php dc", but everything is working fine (pfBlockerNG v2.1.1_4) so no need for it, as in ….if it ain't broke, than don't try to fix it ;)

Do you have any MaxMind update errors in /var/log/pfblockerng/geoip.log?

I would manually run the update and see if it reports any errors…
php /usr/local/www/pfblockerng/pfblockerng.php dc

I encoutered no errors, so no log file. Btw I am on a APU2C4, 2.3.2-RELEASE-p1 (amd64) and pfBlockerNG 2.1.1_4

gcu_greyarea

pfBlockerNG on Bridge

Just wondering if anybody else has seen this.

I run a Bridged Setup (WAN-LAN) and have setup the Management IP on the Bridge Interface (192.168.15.215). This Interface also listens on other ports (OpenVPN , NTP) including 53 for unbound.

DNSBL listens on the default ports and the VIP is 192.168.15.216. I know that ideally this would be a different net, but DNSBL appears to work fine.

I do get strange DNSBL Log entries which I do not understand:
e.g.
DNSBL Recject,Oct 22 16:09:42,192.168.15.216,192.168.15.1, | / |

The pfBlcokerNG Alerts Tab shows the same entry with "No Match" . i.e. unable to determine which DNSBL Feed triggered the "Reject"

Ip Address 192.168.15.1 is the gateway (an ISP Provided router I cannot replace). I don't know why 192.168.15.1 would want to contact the DNSBL VIP?

gcu_greyarea

I was able to narrow down the issue through further testing.

The log entry and its assorted Alert ("No Match") …

DNSBL Recject,Oct 22 16:09:42,192.168.15.216,192.168.15.1, | / |

... can be reproduced by manually testing the DNSBL VIP Ports. e.g. http://vip-ip:80 (8081) . The error is generated for both https and https on ports (80, 8081, 443, 8433)

As it turns out our ISP Provided router wants to connect to the DNSBL VIP IP via port 80 (Seen via packet capture)...
Since WAN and LAN Interfaces are Bridged and filtering is done on the Bridge IF I have no way of blocking the Routers connection attempts...

blueduckdock

Quick Q- in the sync section there is "Sync to configured backup server." Is this for CARP installs?
I'd like to have my HQ box sync not only to the other carp member but also our regional offices. Wondering if/how this works.