分类： LINUX
2006-10-26 13:53:58
ulogd-2.x, the second generation
Netfilter Userspace logging daemon. ulogd makes use of the Linux >= 2.6.14
nfnetlink_log subsystem, but also provides backwards compatibility for Linux
>= 2.4.0 ipt_ULOG.
ulogd-2.x wants to provide a flexible, almost universal logging daemon for netfilter logging. This encompasses both packet-based logging (logging of policy violations) and flow-based logging, e.g. for accounting purpose.
ulogd consists of a small core and a number of plugins. All the real power lies in the plugins, and in the user who configures the interactions between those plugins.
Input plugins acts data source. They get data from somewhere outside of ulogd, and convert it into a list of ulogd keys.
Filter plugins interpret and/or filter data that was received from the Input Plugin. A good example is parsing a raw packet into IPv4 / TCP / ... header information.
Output plugins describe how and where to put the information gained by the Input Plugin and processed by one or more Filter Plugins. The easiest way is to build a line per packet and fprint it to a file. Some people might want to log into a SQL database or want an output conforming to the IETF IPFIX language.
By means of the configuration file, the administrator can build any number of Plugin Stacks. A plugin stack is a seris of plugins, starting with an Input plugin, none, one or multiple filter plugins, and one output plugin on top.
The major clue is providing a framework which is as flexible as possible. Nobody knows what strange network protocols are out there :) But at the same time, logging of a packet filter is often very performance critical. Like all ulogd releases since 0.3.x, the ulogd-2.x core doesn't do any dynamic allocations at runtime. Yes, obviously, at startup time the config file is parsed, and allocations are made. But after that, everything is pre-allocated. As an additional improvement over ulogd-1.x, there are also no hashtable lookups for key resolval. All input/output keys of plugins within every stakc are resolved at config file parsing time, and directly interconnected by pointers.
To use the NFCT or NFLOG input plugin, you will need a 2.6.14 or later kernel. For old-style ULOG logging, you need a kernel >= 2.4.18.
Download the ulogd package from and untar it.
If you want to build ulogd with MySQL support, type './configure --with-mysql'. You may also have to specify the path of the mysql libraries using '--with-mysql=path'. To build ulogd without MySQL support, just use './configure'.
To compile and install the program, call 'make install'.
I also provide a SRPM, which should compile on almost any rpm-based distribution. It is available at
Just download the package and do the usual 'rpm --rebuild
Just add rules using the ULOG target to your firewalling chain. A very basic example:
To increase logging performance, try to use the
option (where 1 < N <= 50). The number you specify is the amount of packets batched together in one multipart netlink message. If you set this to 20, the kernel schedules ulogd only once every 20 packets. All 20 packets are then processed by ulogd. This reduces the number of context switches between kernel and userspace.
Of course you can combine the ULOG target with the different netfilter match modules. For a more detailed description, have a look at the netfilter HOWTO's, available on the netfilter homepage.
The number of the netlink multicast group to which ULOG'ed packets are sent. You will have to use the same group number in the ULOG target and ulogd in order to make logging work.
Copyrange. This works like the 'snaplen' parameter of tcpdump. You can specify
a number of bytes up to which the packet is copied. If you say '40', you will
receive the first fourty bytes of every packet. Leave it to 0
Queue threshold. If a packet is matched by the iptables rule, and already N packets are in the queue, the queue is flushed to userspace. You can use this to implement a policy like: Use a big queue in order to gain high performance, but still have certain packets logged immediately to userspace.
A string that is associated with every packet logged by this rule. You can use this option to later tell from which rule the packet was logged.
The ipt_ULOG kernel module has a couple of module loadtime parameters which can (and should) be tuned to accomodate the needs of the application:
Netlink buffer size. A buffer of the specified size N is allocated for every netlink group that is used. Please note that due to restrictions of the kernel memory allocator, we cannot have a buffer size > 128kBytes. Larger buffer sizes increase the performance, since less kernel/userspace context switches are needed for the same amount of packets. The backside of this performance gain is a potentially larger delay. The default value is 4096 bytes, which is quite small.
The flushtimeout determines, after how many clock ticks (on alpha: 1ms, on x86 and most other platforms: 10ms time units) the buffer/queue is to be flushed, even if it is not full. This can be used to have the advantage of a large buffer, but still a finite maximum delay introduced. The default value is set to 10 seconds.
This would use a buffer size of 64k and a flushtimeout of 100 clockticks (1 second on x86).
ulogd is what this is all about, so let's describe it's configuration...
All configurable parameters of ulogd are in the configfile, typically located at '/etc/ulogd.conf'.
The following configuration parameters are available:
The netlink multicast group, which ulgogd should bind to. This is the same as given with the '--ulog-nlgroup' option to iptables.
The main logfile, where ulogd reports any errors, warnings and other unexpected conditions. Apart from a regular filename, the following special values can be used; ``syslog'' to log via the unix syslog(3) mechanism. ``stdout'' to log to stdout.
This specifies, how verbose the logging to logfile is. Currently defined loglevels are: 1=debug information, 3=informational messages, 5=noticable exceptional conditions, 7=error conditions, 8=fatal errors, program abort.
This option is followed by a filename of a ulogd plugin, which ulogd shold load upon initialization. This option may appear more than once.
Size of the netlink socket receive memory. You should set this to at least the size of the kernel buffer (nlbufsiz parameter of the ipt_ULOG module). Please note that there is a maximum limit in /proc/sys/net/core/rmem_max which you cannot exceed by increasing the ``rmem'' parameter. You may need to raise the system-wide maximum limit before.
Size of the receive buffer. You should set this to at least the socket receive buffer (rmem).
Apart from the configfile, there are a couple of commandline options to ulogd:
Print a help message about the commandline options.
Print version information about ulogd.
For off into daemon mode. Unless you are debugging, you will want to use this most of the time.
Using this commandline option, an alternate config file can be used. This is important if multiple instances of ulogd are to be run on a single machine.
ulogd understands two kinds of signals:
Close and re-open all logfiles. This is mainly intended for logrotate scripts. Also closes and re-opens database connections.
Reload configuration file. This is not fully implemented yet.
It is important to understand that ulogd without plugins does nothing. It will receive packets, and do nothing with them.
There are two kinds of plugins, interpreter and output plugins. Interpreter plugins parse the packet, output plugins write the interpreted information to some logfile/database/...
ulogd comes with the following input plugins:
The good old ipt_ULOG input plugin. This basically emulates ulogd-1.x which didn't have input plugins.
This interfaces the new nfnetlink_log interface. To compile, you need libnetfilter_log installed in your system.
This interfaces the nfnetlink_conntrack kernel subsystem, and provides flow-based logging. To compile, you need libnetfilter_conntrack installed on your system.
ulogd comes with the following interpreter plugins:
Basic interpreter plugin for nfmark, timestamp, mac address, ip header, tcp header, udp header, icmp header, ah/esp header... Most people will want to load this very important plugin.
Example interpreter plugin to log plaintext passwords as used with FTP and POP3. Don't blame me for writing this plugin! The protocols are inherently insecure, and there are a lot of other tools for sniffing passwords... it's just an example.
Filter plugin that provides translation from the numerical ifindex (e.g. '1') to the network interface name (e.g. 'eth4').
This is a 'virtual interpreter'. It doesn't really return any information on the packet itself, rather the local system time and hostname. Please note that the time is the time at the time of logging, not the packets receive time.
ulogd comes with the following output plugins:
A very simple output module, dumping all packets in the format
to a file. The only useful application is debugging.
key=value
key=value
...
===>PACKET BOUNDARY
...
The module defines the following configuration directives:
The filename where it should log to. The default is
/var/log/ulogd.pktlog
An output module which tries to emulate the old syslog-based LOG targed as far as possible. Logging is done to a seperate textfile instead of syslog, though.
The module defines the following configuration directives:
The filename where it should log to. The default is
/var/log/ulogd.syslogemu
Set this to 1 if you want to have your logfile written
synchronously. This may reduce performance, but makes your log-lines appear
immediately. The default is 0
An output plugin for logging into a mysql database. This is only compiled if you have the mysql libraries installed, and the configure script was able to detect them. (that is: --with-mysql was specified for ./configure)
The plugin automagically inserts the data into the configured table; It connects to mysql during the startup phase of ulogd and obtains a list of the columns in the table. Then it tries to resolve the column names against keys of interpreter plugins. This way you can easily select which information you want to log - just by the layout of the table.
If, for example, your table contains a field called 'ip_saddr', ulogd will resolve this against the key 'ip.saddr' and put the ip address as 32bit unsigned integer into the table.
You may want to have a look at the file 'doc/mysql.table' as an
example table including fields to log all keys from ulogd_BASE.so. Just delete
the fields you are not interested in, and create the table.
The module defines the following configuration directives:
Name of the table to which ulogd should log.
Name of the mysql database.
Name of the mysql database host.
TCP port number of mysql database server.
Name of the mysql user.
Password for mysql.
An output plugin for logging into a postgresql database. This is only compiled if you have the mysql libraries installed, and the configure script was able to detect them. (that is: --with-pgsql was specified for ./configure)
The plugin automagically inserts the data into the configured table; It connects to pgsql during the startup phase of ulogd and obtains a list of the columns in the table. Then it tries to resolve the column names against keys of interpreter plugins. This way you can easily select which information you want to log - just by the layout of the table.
If, for example, your table contains a field called 'ip_saddr', ulogd will resolve this against the key 'ip.saddr' and put the ip address as 32bit unsigned integer into the table.
You may want to have a look at the file 'doc/mysql.table' as an
example table including fields to log all keys from ulogd_BASE.so. Just delete
the fields you are not interested in, and create the table.
The module defines the following configuration directives:
Name of the table to which ulogd should log.
Name of the database.
Name of the mysql database host.
TCP port number of database server.
Name of the sql user.
Password for sql user.
An output plugin that can be used to generate libpcap-style packet logfiles. This can be useful for later analysing the packet log with tools like tcpdump or ethereal.
The module defines the following configuration directives:
The filename where it should log to. The default is:
/var/log/ulogd.pcap
Set this to 1 if you want to have your pcap logfile written
synchronously. This may reduce performance, but makes your packets appear
immediately in the file on disk. The default is 0
An output plugin for logging into a SQLITE v3 database. This is only compiled if you have the sqlite libraries installed, and the configure script was able to detect them. (that is: --with-sqlite3 was specified for ./configure)
The plugin automagically inserts the data into the configured table; It opens the sqlite db during the startup phase of ulogd and obtains a list of the columns in the table. Then it tries to resolve the column names against keys of interpreter plugins. This way you can easily select which information you want to log - just by the layout of the table.
If, for example, your table contains a field called 'ip_saddr', ulogd will resolve this against the key 'ip.saddr' and put the ip address as 32bit unsigned integer into the table.
You may want to have a look at the file 'doc/sqlite3.table' as an
example table including fields to log all keys from ulogd_BASE.so. Just delete
the fields you are not interested in, and create the table.
The module defines the following configuration directives:
Name of the table to which ulogd should log.
Name of the database.
Size of the sqlite buffer.
An output plugin that really logs via syslogd. Lines will look exactly like printed with traditional LOG target.
The module defines the following configuration directives:
The syslog facility (LOG_DAEMON, LOG_KERN, LOG_LOCAL0 .. LOG_LOCAL7, LOG_USER)
The syslog level (LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR, LOG_WARNING, LOG_NOTICE, LOG_INFO, LOG_DEBUG)
All comments / questions / ... are appreciated.
Just drop me a note to laforge@gnumonks.org
Please note also that there is now a mailinglist, ulogd@lists.gnumonks.org. You can subscribe at .
The preferred method for reporting bugs is the netfilter bugzilla system, available at .