ELK + pfSense 2.3 Working

[ronv]

Hi all,

trying to get this going with PFsense 2.3.4 and ELK 5.4 - all components are talking ok, and I can get the JSON Dashboard, Search and Visualization up and running - almost…:

• when I import the visualizations, Kibana complains that the tags geoip.country_name and geoip.city_name are not available.
• I checked 11-pfsense.conf (which I used from this site) against the spec at https://www.elastic.co/guide/en/logstash/current/plugins-filters-geoip.html, and there does not appear to be any issue with this - that is, it looks like those tags should be returned.

Anything else I could check, or logs I could provide?

kind regards

Ron

[hamed_forum]

the log send from pfsense where is save on elk?
i change the elk server and how to export import log on prvise server?

[pfBasic]

Any differences to get this running on 2.4.0 BETA?

[pfBasic]

I finally got this up & running on pfSense 2.4.0 BETA with the help of AR15USR and some people on IRC.

Initially I was having trouble getting the Index Patterns to populate in the first step of Kibana. I had followed doktornotor's advice for setting up MaxMind. For whatever reason that didn't work for me so I just did it according to http://pfelk.3ilson.com/ and it worked.

Next, I had everything stable and logs being imported, but all logs were being tagged "_grokparsefailure" & "_geoip_lookup_failure" and since the pattern wasn't matching, it wasn't putting out any useful fields/information. This was also preventing me from importing the Visualizations.json due to not having the applicable fields available.

After way too much time troubleshooting and trying to figure out what was happening and why I was given some direction and pointed to the grok debugger by a kind IRC user. https://grokdebug.herokuapp.com/
For anyone looking to troubleshoot or modify their own grok pattern files, here's what I could make of the fields in 2.4.0 BETA's Rsyslog format. https://forum.pfsense.org/index.php?topic=133354.msg733494#msg733494
Run a pcap to see exactly what your pfSense box is sending to your ELK server.

It turned out that all I needed to do was change one character in /etc/logstash/conf.d/patterns/pfsense2-3.grok and reboot.

I changed line 16 (PFSENSE_LOG_DATA)
From:

PFSENSE_LOG_DATA (%{INT:rule}),(%{INT:sub_rule}),,(%{INT:tracker}),(%{WORD:iface}),(%{WORD:reason}),(%{WORD:action}),(%{WORD:direction}),(%{INT:ip_ver}),

To:

PFSENSE_LOG_DATA (%{INT:rule}),(%{INT:sub_rule})?,,(%{INT:tracker}),(%{WORD:iface}),(%{WORD:reason}),(%{WORD:action}),(%{WORD:direction}),(%{INT:ip_ver}),

That's it, one "?".

After that, log files were parsing successfully, I refreshed my Index Pattern Field List to pull in all of the new fields, imported the Visualizations.json and opened up the Dashboard. All is working now on my single core atom with 2GB DDR2!

@doktornotor:

@johnpoz:

I have it running, but elasticstack doesn't seem to want to stay running.  Haven't had time to look into why.

Make sure you've allocated at least 4GiB of RAM to this thing. (Java  >:( ::))

I have this up and running (for home use) on an old netbook with an atom N450 (Pineview ~2010, single core 1.66GHz) with 2GB DDR2. I had to significantly lower RAM usage in the following two files to get it working. Currently using <1.5GB RAM, the OS is lubuntu with GUI service disabled. It's also running a Unifi controller. Dashboard is slow to load even for a small home network but it works! I couldn't justify buying anything to get an ELK stack for my home network.

/etc/elasticsearch/jvm.options

/etc/logstash/jvm.options

[idealanthony]

@BrunoCAVILLE:

Eveything works well except the maps visualization, someone can help?

@BrunoCAVILLE - I'm having the same problem as you did.  I used the revised visualization file due to the .keyword issue.  I've attempted to merge back in the country sections from the http://pfelk.3ilson.com/ visualization file, but still no luck.  Just wanted to know if you were able to identify/ resolve the issue?

https://forum.pfsense.org/index.php?topic=125376.0

[pfBasic]

Did you refresh your fields list (Management / Index Patterns) after a number of your log files were successfully parsed?

If not, do that first then try to import the pf3lk visulization.json.

The import fails if you don't have the appropriate fields available.

[8ayM]

I just wanted to add that the Kibana4 init script from the OP is no longer listed via a link as others were so I wanted to copy it here in text form as it did take me a moment to realize that the scripts were all included in a zip file in the op as well.

Kibana4 init script:

#!/bin/sh

/etc/init.d/kibana4 – startup script for kibana4

bsmith@the408.com 2015-02-20; used elasticsearch init script as template

https://github.com/akabdog/scripts/edit/master/kibana4_init

BEGIN INIT INFO

Provides:          kibana4

Required-Start:    $network $remote_fs $named

Required-Stop:    $network $remote_fs $named

Default-Start:    2 3 4 5

Default-Stop:      0 1 6

Short-Description: Starts kibana4

Description:      Starts kibana4 using start-stop-daemon

END INIT INFO

#configure this with wherever you unpacked kibana:
KIBANA_BIN=/opt/kibana4/bin

PID_FILE=/var/run/$NAME.pid
PATH=/bin:/usr/bin:/sbin:/usr/sbin:$KIBANA_BIN
DAEMON=$KIBANA_BIN/kibana
NAME=kibana4
DESC="Kibana4"

if [ id -u -ne 0 ]; then
        echo "You need root privileges to run this script"
        exit 1
fi

. /lib/lsb/init-functions

if [ -r /etc/default/rcS ]; then
        . /etc/default/rcS
fi

case "$1" in
  start)
        log_daemon_msg "Starting $DESC"

pid=pidofproc -p $PID_FILE kibana
        if [ -n "$pid" ] ; then
                log_begin_msg "Already running."
                log_end_msg 0
                exit 0
        fi

# Start Daemon
        start-stop-daemon –start --pidfile "$PID_FILE" --make-pidfile --background --exec $DAEMON
        log_end_msg $?
        ;;
  stop)
        log_daemon_msg "Stopping $DESC"

if [ -f "$PID_FILE" ]; then
                start-stop-daemon –stop --pidfile "$PID_FILE"
                        --retry=TERM/20/KILL/5 >/dev/null
                if [ $? -eq 1 ]; then
                        log_progress_msg "$DESC is not running but pid file exists, cleaning up"
                elif [ $? -eq 3 ]; then
                        PID="cat $PID_FILE"
                        log_failure_msg "Failed to stop $DESC (pid $PID)"
                        exit 1
                fi
                rm -f "$PID_FILE"
        else
                log_progress_msg "(not running)"
        fi
        log_end_msg 0
        ;;
  status)
        status_of_proc -p $PID_FILE kibana kibana && exit 0 || exit $?
    ;;
  restart|force-reload)
        if [ -f "$PID_FILE" ]; then
                $0 stop
                sleep 1
        fi
        $0 start
        ;;
  *)
        log_success_msg "Usage: $0 {start|stop|restart|force-reload|status}"
        exit 1
        ;;
esac

[8ayM]

@AR15USR:

ando1, any idea what is going on?

PS I ran everyone of your troubleshooting commands and they all error out fyi…

How did you make out with this? I'm running into the same issue and it doesn't look like there was a resolution to the issue.

[AR15USR]

I'm starting to get this error when my Dashboard refreshes: "Courier Fetch: 28 of 325 shards failed."

I've noticed that I'm seeing yellow health and replications:

health status index               uuid                   pri rep docs.count docs.deleted store.size pri.store.size
green  open   logstash-2017.09.09 yhdtjrKHQVycMOCfBmssWQ   5   0     347962            0    150.8mb        150.8mb
green  open   logstash-2017.09.10 Yg98wyN5SYav2dnc8OmFxA   5   0     359406            0      158mb          158mb
green  open   logstash-2017.09.11 mG66BkrDQOSnyJCqI5Ir-w   5   0     380644            0    164.2mb        164.2mb
green  open   logstash-2017.09.12 y26fNsoORtW6cSx1QcE7ZQ   5   0     390537            0    169.2mb        169.2mb
green  open   logstash-2017.09.13 MxyncENMRXqxLnMuJIw1rw   5   0     353464            0    152.2mb        152.2mb
yellow open   logstash-2017.09.14 Gp3dZ-uUTeWv9YIS4calhw   5   1     376975            0    163.5mb        163.5mb
yellow open   logstash-2017.09.15 cq8n4mYYSWGZZrzrb50B-g   5   1     392566            0    165.2mb        165.2mb
yellow open   logstash-2017.09.16 u7aF2fGSSmOJmJCU4odO5w   5   1     210728            0     94.5mb         94.5mb

Anyone know why this is happening all the sudden?

[AR15USR]

Anyone able to give me any clues or places to start with the above?

[tzidore]

Hi

I have been trying to fetch logs from pfSense 2.4.1 all day. I haven't been able to get the grok pattern to work.
The post below does nothing for me. It worked fine on 2.3.4_p1.

Any ideas?

@pfBasic:

I finally got this up & running on pfSense 2.4.0 BETA with the help of AR15USR and some people on IRC.

Initially I was having trouble getting the Index Patterns to populate in the first step of Kibana. I had followed doktornotor's advice for setting up MaxMind. For whatever reason that didn't work for me so I just did it according to http://pfelk.3ilson.com/ and it worked.

Next, I had everything stable and logs being imported, but all logs were being tagged "_grokparsefailure" & "_geoip_lookup_failure" and since the pattern wasn't matching, it wasn't putting out any useful fields/information. This was also preventing me from importing the Visualizations.json due to not having the applicable fields available.

After way too much time troubleshooting and trying to figure out what was happening and why I was given some direction and pointed to the grok debugger by a kind IRC user. https://grokdebug.herokuapp.com/
For anyone looking to troubleshoot or modify their own grok pattern files, here's what I could make of the fields in 2.4.0 BETA's Rsyslog format. https://forum.pfsense.org/index.php?topic=133354.msg733494#msg733494
Run a pcap to see exactly what your pfSense box is sending to your ELK server.

It turned out that all I needed to do was change one character in /etc/logstash/conf.d/patterns/pfsense2-3.grok and reboot.

I changed line 16 (PFSENSE_LOG_DATA)
From:

PFSENSE_LOG_DATA (%{INT:rule}),(%{INT:sub_rule}),,(%{INT:tracker}),(%{WORD:iface}),(%{WORD:reason}),(%{WORD:action}),(%{WORD:direction}),(%{INT:ip_ver}),

To:

PFSENSE_LOG_DATA (%{INT:rule}),(%{INT:sub_rule})?,,(%{INT:tracker}),(%{WORD:iface}),(%{WORD:reason}),(%{WORD:action}),(%{WORD:direction}),(%{INT:ip_ver}),

That's it, one "?".

After that, log files were parsing successfully, I refreshed my Index Pattern Field List to pull in all of the new fields, imported the Visualizations.json and opened up the Dashboard. All is working now on my single core atom with 2GB DDR2!

@doktornotor:

@johnpoz:

I have it running, but elasticstack doesn't seem to want to stay running.  Haven't had time to look into why.

Make sure you've allocated at least 4GiB of RAM to this thing. (Java  >:( ::))

I have this up and running (for home use) on an old netbook with an atom N450 (Pineview ~2010, single core 1.66GHz) with 2GB DDR2. I had to significantly lower RAM usage in the following two files to get it working. Currently using <1.5GB RAM, the OS is lubuntu with GUI service disabled. It's also running a Unifi controller. Dashboard is slow to load even for a small home network but it works! I couldn't justify buying anything to get an ELK stack for my home network.

/etc/elasticsearch/jvm.options

/etc/logstash/jvm.options

[f11]

So after messing about with the Grok debugger, modifying the .grok file to the following seemed to work:

PFSENSE_LOG_DATA (%{INT:rule}),,,(%{INT:tracker}),(%{WORD:iface}),(%{WORD:reason}),(%{WORD:action}),(%{WORD:direction}),(%{INT:ip_ver}),
PFSENSE_IP_SPECIFIC_DATA (%{PFSENSE_IPv4_SPECIFIC_DATA}|%{PFSENSE_IPv6_SPECIFIC_DATA})
PFSENSE_IPv4_SPECIFIC_DATA (%{BASE16NUM:tos}),,(%{INT:ttl}),(%{INT:id}),(%{INT:offset}),(%{WORD:flags}),(%{INT:proto_id}),(%{WORD:proto}),
PFSENSE_IPv4_SPECIFIC_DATA_ECN (%{BASE16NUM:tos}),(%{INT:ecn}),(%{INT:ttl}),(%{INT:id}),(%{INT:offset}),(%{WORD:flags}),(%{INT:proto_id}),(%{WORD:proto}),
PFSENSE_IPv6_SPECIFIC_DATA (%{BASE16NUM:class}),(%{DATA:flow_label}),(%{INT:hop_limit}),(%{WORD:proto}),(%{INT:proto_id}),
PFSENSE_IP_DATA (%{INT:length}),(%{IP:src_ip}),(%{IP:dest_ip}),
PFSENSE_PROTOCOL_DATA (%{PFSENSE_TCP_DATA}|%{PFSENSE_UDP_DATA}|%{PFSENSE_ICMP_DATA}|%{PFSENSE_CARP_DATA})
PFSENSE_TCP_DATA (%{INT:src_port}),(%{INT:dest_port}),(%{INT:data_length}),(%{WORD:tcp_flags}),(%{INT:sequence_number}),(%{INT:ack_number}),(%{INT:tcp_window}),(%{DATA:urg_data}),(%{DATA:tcp_options})
PFSENSE_UDP_DATA (%{INT:src_port}),(%{INT:dest_port}),(%{INT:data_length})
PFSENSE_ICMP_DATA (%{PFSENSE_ICMP_TYPE}%{PFSENSE_ICMP_RESPONSE})
PFSENSE_ICMP_TYPE (?<icmp_type>(request|reply|unreachproto|unreachport|unreach|timeexceed|paramprob|redirect|maskreply|needfrag|tstamp|tstampreply)),
PFSENSE_ICMP_RESPONSE (%{PFSENSE_ICMP_ECHO_REQ_REPLY}|%{PFSENSE_ICMP_UNREACHPORT}| %{PFSENSE_ICMP_UNREACHPROTO}|%{PFSENSE_ICMP_UNREACHABLE}|%{PFSENSE_ICMP_NEED_FLAG}|%{PFSENSE_ICMP_TSTAMP}|%{PFSENSE_ICMP_TSTAMP_REPLY})
PFSENSE_ICMP_ECHO_REQ_REPLY (%{INT:icmp_echo_id}),(%{INT:icmp_echo_sequence})
PFSENSE_ICMP_UNREACHPORT (%{IP:icmp_unreachport_dest_ip}),(%{WORD:icmp_unreachport_protocol}),(%{INT:icmp_unreachport_port})
PFSENSE_ICMP_UNREACHPROTO (%{IP:icmp_unreach_dest_ip}),(%{WORD:icmp_unreachproto_protocol})
PFSENSE_ICMP_UNREACHABLE (%{GREEDYDATA:icmp_unreachable})
PFSENSE_ICMP_NEED_FLAG (%{IP:icmp_need_flag_ip}),(%{INT:icmp_need_flag_mtu})
PFSENSE_ICMP_TSTAMP (%{INT:icmp_tstamp_id}),(%{INT:icmp_tstamp_sequence})
PFSENSE_ICMP_TSTAMP_REPLY (%{INT:icmp_tstamp_reply_id}),(%{INT:icmp_tstamp_reply_sequence}),(%{INT:icmp_tstamp_reply_otime}),(%{INT:icmp_tstamp_reply_rtime}),(%{INT:icmp_tstamp_reply_ttime})

PFSENSE_CARP_DATA (%{WORD:carp_type}),(%{INT:carp_ttl}),(%{INT:carp_vhid}),(%{INT:carp_version}),(%{INT:carp_advbase}),(%{INT:carp_advskew})

DHCPD (%{DHCPDISCOVER}|%{DHCPOFFER}|%{DHCPREQUEST}|%{DHCPACK}|%{DHCPINFORM}|%{DHCPRELEASE})
DHCPDISCOVER %{WORD:dhcp_action} from %{COMMONMAC:dhcp_client_mac}%{SPACE}(\(%{GREEDYDATA:dhcp_client_hostname}\))? via (?<dhcp_client_vlan>[0-9a-z_]*)(: %{GREEDYDATA:dhcp_load_balance})?
DHCPOFFER %{WORD:dhcp_action} on %{IPV4:dhcp_client_ip} to %{COMMONMAC:dhcp_client_mac}%{SPACE}(\(%{GREEDYDATA:dhcp_client_hostname}\))? via (?<dhcp_client_vlan>[0-9a-z_]*)
DHCPREQUEST %{WORD:dhcp_action} for %{IPV4:dhcp_client_ip}%{SPACE}(\(%{IPV4:dhcp_ip_unknown}\))? from %{COMMONMAC:dhcp_client_mac}%{SPACE}(\(%{GREEDYDATA:dhcp_client_hostname}\))? via (?<dhcp_client_vlan>[0-9a-z_]*)(: %{GREEDYDATA:dhcp_request_message})?
DHCPACK %{WORD:dhcp_action} on %{IPV4:dhcp_client_ip} to %{COMMONMAC:dhcp_client_mac}%{SPACE}(\(%{GREEDYDATA:dhcp_client_hostname}\))? via (?<dhcp_client_vlan>[0-9a-z_]*)
DHCPINFORM %{WORD:dhcp_action} from %{IPV4:dhcp_client_ip} via %(?<dhcp_client_vlan>[0-9a-z_]*)
DHCPRELEASE %{WORD:dhcp_action} of %{IPV4:dhcp_client_ip} from %{COMMONMAC:dhcp_client_mac}%{SPACE}(\(%{GREEDYDATA:dhcp_client_hostname}\))? via</dhcp_client_vlan></dhcp_client_vlan></dhcp_client_vlan></dhcp_client_vlan></dhcp_client_vlan></icmp_type>

Essentially the line

PFSENSE_LOG_DATA (%{INT:rule}),(%{INT:sub_rule}),,(%{INT:tracker}),(%{WORD:iface}),(%{WORD:reason}),(%{WORD:action}),(%{WORD:direction}),(%{INT:ip_ver}),

Was changed to

PFSENSE_LOG_DATA (%{INT:rule}),,,(%{INT:tracker}),(%{WORD:iface}),(%{WORD:reason}),(%{WORD:action}),(%{WORD:direction}),(%{INT:ip_ver}),

Now logs seem to be parsing correctly and my dashboard is looking populated again.

[alex_ncus]

NOOB

Anyone been able to get latest ELK running in Docker containers and parsing pfsense 2.4.1 logs? Trying to get pfsense to send log to NAS and then having my Mac workstation running Docker with ELK to parse / analyze data staged on NAS.

Does this even make sense?

[donnydavis]

I used EFK and it works great for me

[vinchi007]

Thanks for sharing! I've setup my ELK and PFSence 2.2 on FreeNAS jail. one thing to note, if you're using FreeNas, logstash startup script will list only logstash.conf file to load so I've combined all of these conf files into one. Also, note that you need to change location path of "patters" location where grok patterns file will reside in.

[AR15USR]

Anyone having issues with GROK failure on 2.4.2? I'm seeing more than 50% failures. Seems to be failing on the inbound block/pass logs, the outbound blocks/pass seem to work fine.

Ive set it up exactly according to http://pfelk.3ilson.com/ and also have setup MaxMind according to Docs Insdructions: https://forum.pfsense.org/index.php?topic=120937.msg671603#msg671603

[thhi]

May this is useful for someone:

We use vlans in our environment and our interface names contain "." (and not only "word" characters). So we modified "PFSENSE_LOG_DATA" in the  grok filter

IFACE \b[a-zA-Z0-9.]+\b
#PFSENSE_LOG_DATA (%{INT:rule}),(%{INT:sub_rule}),,(%{INT:tracker}),(%{WORD:iface}),(%{WORD:reason}),(%{WORD:action}),(%{WORD:direction}),(%{INT:ip_ver}),
PFSENSE_LOG_DATA (%{INT:rule}),(%{INT:sub_rule})?,,(%{INT:tracker}),(%{IFACE:iface}),(%{WORD:reason}),(%{WORD:action}),(%{WORD:direction}),(%{INT:ip_ver}),

[gurulee]

I followed this guide initially: https://www.digitalocean.com/community/tutorials/how-to-install-elasticsearch-logstash-and-kibana-elastic-stack-on-ubuntu-16-04 . At that point I was seeing data from filebeats in Kibana only.

But after merging in these steps, I now do not see any log data from pfsense, or filebeats. Do I need to disable/remove filebeats?

[gurulee]

@gurulee said in ELK + pfSense 2.3 Working:

Kibana

I believe filebeat was screwing things up. I blew away the vm after hours of racking my brain and reinstalling ELK stack. I ended up rebuilding it according to this guide http://pfelk.3ilson.com/2017/10/pfsense-v24xkibanaelasticsearchlogstash.html?m=1
, and it’s working now. I just need to tweak dashboards and visualizations at this point.

[bartkowski]

Have any of you seen https://logz.io offering?