#syslog-ng

Привет, Преди няколко седмици обнових мейл сървъра ми към версия debian 8 jessie. Всичко работеше, но днес забелязах следния проблем във /var/log/mail.log : Jan 26 11:00:57 alpha postfix/smtpd[10642]: warning: premature end-of-input on private/policy while reading input attribute name Jan 26 11:00:58 alpha postfix/smtpd[10642]: warning: premature end-of-input on private/policy while reading input…

After using Loggly for a few years, I stumbled across Logentries (disclosure: referral link). What I was looking for was basically something that is more affordable when the volume increases, and Logentries is a lot more affordable than Loggly as you grow.

If you are just starting out using a remote shipping target (be Loggly, Papertrail or Logentries), it is worth noting that if you simply follow the instructions you will be shipping logs in plain text. Since you’re a smart cookie, you know that’s a very bad idea.

The good news is that most services (at least the three mentioned) do support encryption, but it takes a bit more work. What I did find however was that the instructions for Logentries in particular were pretty bad, so I decided to share this with you to save you the effort it took me to get this working with syslog-ng.

Step 1: Fetch the certificates

This was surprisingly challenging with Logentries. While they openly announce that they support encryption, the links to where you can find said certificates wasn’t as obvious. Long story short, you can find them here. Unfortunately, they do simply give you a URL, but instead encourage you to copy and paste the certs by hand (which is of course far more error prone).

The certificates that you need are the ones under ‘API certificate’.

Go ahead and save the first certificate as api.crt and the intermediate certificate as intermediate.crt.

In theory, you should now verify the hashes, but I wasn’t able to get mine to match, so perhaps they’ve simply forgot to update the hashes. If it’s to any help, this is what I got:

$ md5 *.crt MD5 (api.crt) = 9107ba5545a000ea06cd5fd046102c14 MD5 (intermediate.crt) = 413a2acb5b07cd49cbf916eefcf3ba33

(Please note that this of course will change whenver Logentries updates their certificates.)

With these certificates, we now need to generate a combined certificate that we’ll ship to the nodes. To do this, simply run:

$ cat {intermediate.crt,api.crt} > logentries_full.crt

If you’re lazy, you can also fetch mine from here.

The md5sum for this certificate should be:

$ md5 logentries_full.crt MD5 (logentries_full.crt) = 3918fcd927bb98fa2c23a46e5a4b7820

(Again, this will of course also change whenver Logentries updates their certificates.)

Now create the directory /etc/syslog-ng/keys/ca.d/ on the host and copy logentries_full.crt certificate there.

Step 2: Configure syslog-ng

With the certificate in place, the last step is to create the config file. My preference is to use the conf.d folder for this (which already exists on Debian/Ubuntu installations).

Now create the file /etc/syslog-ng/conf.d/22-logentries.conf and populate it with the following data:

template logentriesTemplate { template("YOURTOKEN $ISODATE $HOST $PROGRAM $MSG\n"); template_escape(no); }; source s_all { unix-stream("/dev/log"); }; destination d_network_logentries { tcp("api.logentries.com" port(20000) tls(peer-verify(required-untrusted) ca_dir("/etc/syslog-ng/keys/ca.d/")) template(logentriesTemplate) ); }; log { source(s_all); destination(d_network_logentries); };

Replace YOURTOKEN with the actual token from Logentries. You also might change some elements in this file to better fit your needs, but what’s important here is the destination-block, which is what is dealing with the TLS aspect.

With the file in place, restart syslog-ng and you should be good to go.

Troubleshooting

If you’re having issues getting this started, the first thing to do is to check the actual logs. There should be a line similar to this if it worked:

Syslog connection established; fd='11', server='AF_INET(a.b.c.d:20000)', local='AF_INET(0.0.0.0:0)'

You might also want to sniff the traffic to verify that the traffic sent out actually is encrypted. You can do that easily with tcpdump as follows:

$ tcpdump -A dst api.logentries.com

Resolved: syslog-ng on Solaris 10 (how to install and configure) #fix #dev #computers

syslog-ng on Solaris 10 (how to install and configure)

If anyone has got syslog-ng working on Solaris 10, do you have a set of instructions I could follow to get it installed and working?

I tried following this http://opensystems.wordpress.com/2006/06/01/replacing-syslog-on-solaris-10-with-syslog-ng/ but got stuck with the syslog-ng service stuck restarting with no indication of why.

Thanks for…

A lot of the logs that are dictated by PCI-DSS for retention simply don't need to be kept for daily review. 

I'm a long time Splunk user and I've grown accustomed to keeping damn near everything there. 

It reaches a certain point that this just isn't feasible. 

Some basic preprocessing with syslog-ng and depositing to a non Splunk location is just the ticket. The following is a perfect example. 

Syslog-NG's filters have all the power and flexibility required to get very granular in what you don't need to review, all the way down to certain Windows event IDs from certain servers. 

It does take a couple tricks for a standard syslog server to act as PCI compliant storage platform - you need to be sure the local access auditing is enabled, and you need to sign the messages as they come in.

My preferred method is using a logrotate script.

As the day's events are archived a single line containing a SHA sum of the events (to prove against tampering, per the PCI standard) is then written to Splunk. 

install syslog-ng using tcp 4800 port aptitude install syslog-ng vi /etc/default/syslog-ng CONSOLE_LOG_LEVEL=1 /etc/init.d/syslog-ng start cd /etc/syslog-ng/ wget http://www.campin.net/syslog-ng/expanded-syslog-ng.conf mv expanded-syslog-ng.conf syslog-ng.conf #---------------------------------------------------------------------- #  Program:  syslog-ng.conf #  Notes:    Embedded most of the manual notes within the configuration #            file.  The original manual can be found at: # #            http://www.balabit.com/products/syslog_ng/reference/book1.html #            http://www.campin.net/syslog-ng/faq.html # #            Many people may find placing all of this information in a #            configuration file a bit redundant, but I have found that #            with a little bit of extra comments and reference, #            maintaining these beasties is much easier. # #            This particular log file was taken from the examples that #            are given at the different web sites, and made to emulate #            the logs of a Mandrake Linux system as much as possible. #            Of course, Unix is Unix, is Linux.  It should be generic #            enough for any Unix system. #---------------------------------------------------------------------- #  16-Mar-03 - REP - Added some extra definitions to the file. #  15-Mar-03 - REP - Added back the comments on filtering. #  27-Feb-03 - REP - Further modified for local environment. #  27-Feb-03 - REP - Updated for new configuration and version 1.6.0 #  12-Dec-02 - REP - Continued updates for writing to databases. #  30-Nov-02 - REP - Initial creation for testing. #---------------------------------------------------------------------- #  Options #---------------------------------------------------------------------- # #  Name                       Values   Description #  -------------------------  -------  ------------------------------------ #  bad_hostname               reg exp  A regexp which matches hostnames #                                      which should not be taken as such. #  chain_hostnames            y/n      Enable or disable the chained #                                      hostname format. #  create_dirs                y/n      Enable or disable directory creation #                                      for destination files. #  dir_group                  groupid #  dir_owner                  userid #  dir_perm                   perm #  dns_cache                  y/n      Enable or disable DNS cache usage. #  dns_cache_expire           num      Number of seconds while a successful #                                      lookup is cached. #  dns_cache_expire_failed    num      Number of seconds while a failed #                                      lookup is cached. #  dns_cache_size             num      Number of hostnames in the DNS cache. #  gc_busy_threshold          num      Sets the threshold value for the #                                      garbage collector, when syslog-ng is #                                      busy. GC phase starts when the number #                                      of allocated objects reach this #                                      number. Default: 3000. #  gc_idle_threshold          num      Sets the threshold value for the #                                      garbage collector, when syslog-ng is #                                      idle. GC phase starts when the number #                                      of allocated objects reach this #                                      number. Default: 100. #  group                      groupid #  keep_hostname              y/n      Enable or disable hostname rewriting. #                                      This means that if the log entry had #                                      been passed through at least one other #                                      logging system, the ORIGINAL hostname #                                      will be kept attached to the log.   #                                      Otherwise the last logger will be #                                      considered the log entry owner and #                                      the log entry will appear to have #                                      come from that host. #  log_fifo_size              num      The number of lines fitting to the #                                      output queue #  log_msg_size               num      Maximum length of message in bytes. #  long_hostnames             on/off   This options appears to only really #                                      have an affect on the local system. #                                      which removes the source of the log. #                                      As an example, normally the local #                                      logs will state src@hostname, but #                                      with this feature off, the source #                                      is not reported. #  mark                       num      The number of seconds between two #                                      MARK lines. NOTE: not implemented #                                      yet. #  owner                      userid #  perm                       perm #  stats                      num      The number of seconds between two #                                      STATS. #  sync                       num      The number of lines buffered before #                                      written to file #  time_reap                  num      The time to wait before an idle #                                      destination file is closed. #  time_reopen                num      The time to wait before a died #                                      connection is reestablished #  use_dns                    y/n      Enable or disable DNS usage. #                                      syslog-ng blocks on DNS queries, #                                      so enabling DNS may lead to a #                                      Denial of Service attack. To #                                      prevent DoS, protect your #                                      syslog-ng network endpoint with #                                      firewall rules, and make sure that #                                      all hosts, which may get to #                                      syslog-ng is resolvable. #  use_fqdn                   y/n      Add Fully Qualified Domain Name #                                      instead of short hostname. #  use_time_recvd             y/n      Use the time a message is #                                      received instead of the one #                                      specified in the message. #---------------------------------------------------------------------- #  15-Mar-03 - REP - Since some of the clocks are not quite right, we #                    are going to go ahead and just use the local time #                    as the master time. #  12-Mar-03 - REP - We have taken a few configuration options from the #                    newer Solaris configuration because some of the #                    reasons are valid for us as well.  We have increased #                    the log_msg_size and log_fifo_size to increase the #                    amount of buffering that we do.  While for most #                    systems this may not have a noticeable affect, it #                    will for systems that are at the end of a lot of #                    logging systems. #  20-Dec-02 - REP - Changed the stat() time from the default of 10 #                    minutes to once an hour. #---------------------------------------------------------------------- options   {     chain_hostnames(no);     create_dirs (no);     dir_perm(0755);     dns_cache(yes);     keep_hostname(yes);     log_fifo_size(2048);     log_msg_size(8192);     long_hostnames(on);     perm(0644);     stats(3600);     sync(0);     time_reopen (10);     use_dns(yes);     use_fqdn(yes);   }; #---------------------------------------------------------------------- #  Sources #---------------------------------------------------------------------- # #  fifo/pipe     - The pipe driver opens a named pipe with the #                  specified name, and listens for messages. It's #                  used as the native message getting protocol on #                  HP-UX. #  file          - Usually the kernel presents its messages in a #                  special file (/dev/kmsg on BSDs, /proc/kmsg on #                  Linux), so to read such special files, you'll need #                  the file() driver. Please note that you can't use #                  this driver to follow a file like tail -f does. #  internal      - All internally generated messages "come" from this #                  special source. If you want warnings, errors and #                  notices from syslog-ng itself, you have to include #                  this source in one of your source statements. #  sun-streams   - Solaris uses its STREAMS API to send messages to #                  the syslogd process. You'll have to compile #                  syslog-ng with this driver compiled in (see #                  ./configure --help). # #                  Newer versions of Solaris (2.5.1 and above), uses a #                  new IPC in addition to STREAMS, called door to #                  confirm delivery of a message. Syslog-ng supports #                  this new IPC mechanism with the door() option. # #                  The sun-streams() driver has a single required #                  argument, specifying the STREAMS device to open and #                  a single option. #  tcp/udp       - These drivers let you receive messages from the #                  network, and as the name of the drivers show, you #                  can use both UDP and TCP as transport. # #                  UDP is a simple datagram oriented protocol, which #                  provides "best effort service" to transfer #                  messages between hosts. It may lose messages, and #                  no attempt is made to retransmit such lost #                  messages at the protocol level. # #                  TCP provides connection-oriented service, which #                  basically means a flow-controlled message pipeline. #                  In this pipeline, each message is acknowledged, and #                  retransmission is done for lost packets. Generally #                  it's safer to use TCP, because lost connections can #                  be detected, and no messages get lost, but #                  traditionally the syslog protocol uses UDP. # #                  None of tcp() and udp() drivers require positional #                  parameters. By default they bind to 0.0.0.0:514, #                  which means that syslog-ng will listen on all #                  available interfaces, port 514. To limit accepted #                  connections to one interface only, use the #                  localip() parameter as described below. # #    Options: # #    Name            Type    Description                       Default #    --------------  ------  --------------------------------  -------- #    ip or local ip  string  The IP address to bind to. Note   0.0.0.0 #                            that this is not the address #                            where messages are accepted #                            from. #    keep-alive     y/n      Available for tcp() only, and     yes #                            specifies whether to close #                            connections upon the receival #                            of a SIGHUP signal. #    max-connections number  Specifies the maximum number of   10 #                            simultaneous connections. #    port or local   port    number The port number to bind     514 #                            to. #    --------------  ------  --------------------------------  -------- # #  unix-stream   -  unix-dgram - These two drivers behave similarly: #                   they open the given AF_UNIX socket, and start #                   listening on them for messages. unix-stream() is #                   primarily used on Linux, and uses SOCK_STREAM #                   semantics (connection oriented, no messages are #                   lost), unix-dgram() is used on BSDs, and uses #                   SOCK_DGRAM semantics, this may result in lost #                   local messages, if the system is overloaded. # #                   To avoid denial of service attacks when using #                   connection-oriented protocols, the number of #                   simultaneously accepted connections should be #                   limited. This can be achieved using the #                   max-connections() parameter. The default value of #                   this parameter is quite strict, you might have to #                   increase it on a busy system. # #                   Both unix-stream and unix-dgram has a single #                   required positional argument, specifying the #                   filename of the socket to create, and several #                   optional parameters. # #    Options: # #    Name            Type    Description                       Default #    --------------  ------  --------------------------------  -------- #    group           string  Set the gid of the socket.        root #    keep-alive      y/n     Selects whether to keep           yes #                            connections opened when #                            syslog-ng is restarted, can be #                            used only with unix-stream(). #    max-connections numb    Limits the number of              10 #                            simultaneously opened #                            connections. Can be used only #                            with unix-stream(). #    owner           string  Set the uid of the socket.        root #    perm            num     Set the permission mask. For      0666 #                            octal numbers prefix the number #                            with '0', e.g. use 0755 for #                            rwxr-xr-x. #---------------------------------------------------------------------- #  Notes:    For Linux systems (and especially RedHat derivatives), #            they have a second logging process for kernel messages.   #            This source is /proc/kmsg.  If you are running this on a #            system that is not Linux, then the source entry for this #            should be removed. # #            It seems that there is some performance questions related #            to what type of source stream should be used for Linux #            boxes.  The documentation states the /dev/log should use #            unix-stream, but from the mailing list it has been #            strongly suggested that unix-dgram be used. # #  WARNING:  TCP wrappers has been enabled for this system, and unless #            you also place entries in /etc/hosts.allow for each of the #            devices that will be delivering logs via TCP, you will #            NOT receive the logs. # #            Also note that if there is any form of a local firewall, #            this will also need to be altered such that the incoming #            and possibly outgoing packets are allowed by the firewall #            rules. #---------------------------------------------------------------------- #  There has been a lot of debate on whether everything should be put #  to a single source, or breakdown all the sources into individual #  streams.  The greatest flexibility would be in many, but the most #  simple is the single.  Since we wrote this file, we have chosen the #  route of maximum flexibility. # #  For those of you that like simplicity, this could have also been #  done as the follows: # #  source src #    { #      internal(); #      pipe("/proc/kmsg" log_prefix("kernel: ")); #      tcp(ip(127.0.0.1) port(4800) keep-alive(yes)); #      udp(); #      unix-stream("/dev/log"); #    }; # #  You would also have to change all the log statements to only #  reference the now single source stream. #---------------------------------------------------------------------- #  16-Mar-03 - REP - The default number of allowed TCP connects is set #                    very low for a logserver.  This value should only #                    be set greater than the default for servers that #                    will actually be serving that many systems. #---------------------------------------------------------------------- source s_dgram   { unix-dgram("/dev/log"); }; source s_internal   { internal(); }; source s_kernel   { pipe("/proc/kmsg" log_prefix("kernel: ")); }; source s_tcp   { tcp(port(4800) keep-alive(yes) max_connections(100)); }; #---------------------------------------------------------------------- #  Destinations #---------------------------------------------------------------------- # #  fifo/pipe    - This driver sends messages to a named pipe like #                 /dev/xconsole # #                 The pipe driver has a single required parameter, #                 specifying the filename of the pipe to open, and #                 no options. #  file         - The file driver is one of the most important #                 destination drivers in syslog-ng. It allows you to #                 output messages to the named file, or as you'll see #                 to a set of files. # #                 The destination filename may include macros which #                 gets expanded when the message is written, thus a #                 simple file() driver may result in several files #                 to be created. Macros can be included by prefixing #                 the macro name with a '$' sign (without the quotes), #                 just like in Perl/PHP. # #                 If the expanded filename refers to a directory #                 which doesn't exist, it will be created depending #                 on the create_dirs() setting (both global and a per #                 destination option) # #                 WARNING: since the state of each created file must #                 be tracked by syslog-ng, it consumes some memory #                 for each file. If no new messages are written to a #                 file within 60 seconds (controlled by the time_reap #                 global option), it's closed, and its state is freed. # #                 Exploiting this, a DoS attack can be mounted against #                 your system. If the number of possible destination #                 files and its needed memory is more than the amount #                 your logserver has. # #                 The most suspicious macro is $PROGRAM, where the #                 possible variations is quite high, so in untrusted #                 environments $PROGRAM usage should be avoided. # #    Macros: # #    Name               Description #    -----------------  ----------------------------------------------- #    DATE               Date of the transaction. #    DAY                The day of month the message was sent. #    FACILITY           The name of the facility, the message is tagged #                       as coming from. #    FULLDATE           Long form of the date of the transaction. #    FULLHOST           Full hostname of the system that sent the log. #    HOST               The name of the source host where the message #                       is originated from. If the message traverses #                       several hosts, and chain_hostnames() is on, #                       the first one is used. #    HOUR               The hour of day the message was sent. #    ISODATE            Date in ISO format. #    MIN                The minute the message was sent. #    MONTH              The month the message was sent. #    MSG or MESSAGE     Message contents. #    PRIORITY or LEVEL  The priority of the message. #    PROGRAM            The name of the program the message was sent by. #    SEC                The second the message was sent. #    TAG                The priority and facility encoded as a 2 digit #                       hexadecimal number. #    TZ                  The time zone or name or abbreviation. e.g. 'PDT' #    TZOFFSET           The time-zone as hour offset from GMT. e.g. #                       '-0700' #    WEEKDAY            The 3-letter name of the day of week the #                       message was sent, e.g. 'Thu'. #    YEAR               The year the message was sent. Time expansion #                       macros can either use the time specified in #                       the log message, e.g. the time the log message #                       is sent, or the time the message was received #                       by the log server. This is controlled by the #                       use_time_recvd() option. #    -----------------  ----------------------------------------------- # #    Options: # #    Name            Type    Description                       Default #    --------------  ------  --------------------------------  -------- #    compress        y/n     Compress the resulting logfile    global #                            using zlib. NOTE: this is not     setting #                            implemented as of 1.3.14. #    reate_dirs      y/n     Enable creating non-existing      no #                            directories. #    dir_perm        num     The permission mask of            0600 #                            directories created by #                            syslog-ng. Log directories are #                            only created if a file after #                            macro expansion refers to a #                            non-existing directory, and dir #                            creation is enabled using #                            create_dirs(). #    encrypt         y/n     Encrypt the resulting file.       global #                            NOTE: this is not implemented as  setting #                            of 1.3.14. #    fsync           y/n     Forces an fsync() call on the #                            destination fd after each write. #                            Note: this may degrade #                            performance seriously #    group           string  Set the group of the created      root #                            filename to the one specified. #    log_fifo_size   num     The number of entries in the      global #                            output fifo.                      setting #    owner           string  Set the owner of the created      root #                            filename to the one specified. #    perm            num     The permission mask of the file   0600 #                            if it is created by syslog-ng. #    remove_if_older num     If set to a value higher than 0,  0 #                            before writing to a file, #                            syslog-ng checks whether this #                            file is older than the specified #                            amount of time (specified in #                            seconds). If so, it removes the #                            existing file and the line to #                            be written is the first line in #                            a new file with the same name. #                            In combination with e.g. the #                            $WEEKDAY macro, this is can be #                            used for simple log rotation, #                            in case not all history need to #                            be kept. #    sync_freq       num     The logfile is synced when this   global #                            number of messages has been       setting #                            written to it. #    template        string  Specifies a template which #                            specifies the logformat to be #                            used in this file. The possible #                            macros are the same as in #                            destination filenames. #    template_escape y/n     Turns on escaping ' and " in      yes #                            templated output files. It is #                            useful for generating SQL #                            statements and quoting string #                            contents so that parts of your #                            log message don't get #                            interpreted as commands to the #                            SQL server. #    --------------  ------  --------------------------------  -------- # #  program      - This driver fork()'s executes the given program with #                 the given arguments and sends messages down to the #                 stdin of the child. # #                 The program driver has a single required parameter, #                 specifying a program name to start and no options. #                 The program is executed with the help of the current #                 shell, so the command may include both file patterns #                 and I/O redirection, they will be processed. # #                 NOTE: the program is executed once at startup, and #                 kept running until SIGHUP or exit. The reason is to #                 prevent starting up a large number of programs for #                 messages, which would imply an easy DoS. #  tcp/udp      - This driver sends messages to another host on the #                 local intranet or internet using either UDP or TCP #                 protocol. # #                 Both drivers have a single required argument #                 specifying the destination host address, where #                 messages should be sent, and several optional #                 parameters. Note that this differs from source #                 drivers, where local bind address is implied, and #                 none of the parameters are required. # #    Options: # #    Name            Type    Description                       Default #    --------------  ------  --------------------------------  -------- #    localip         string  The IP address to bind to before  0.0.0.0 #                            connecting to target. #    localport       num     The port number to bind to.       0 #    port/destport   num     The port number to connect to.   514 #    --------------  ------  --------------------------------  -------- #  usertty      - This driver writes messages to the terminal of a #                 logged-in user. # #                 The usertty driver has a single required argument, #                 specifying a username who should receive a copy of #                 matching messages, and no optional arguments. #  unix-dgram   - unix-stream -  This driver sends messages to a unix #                 socket in either SOCK_STREAM or SOCK_DGRAM mode. # #                 Both drivers have a single required argument #                 specifying the name of the socket to connect to, and #                 no optional arguments. #---------------------------------------------------------------------- #---------------------------------------------------------------------- #  Standard Log file locations #---------------------------------------------------------------------- destination authlog        { file("/var/log/auth.log"); }; destination bootlog        { file("/var/log/boot.log"); }; destination debug          { file("/var/log/debug"); }; destination explan         { file("/var/log/explanations"); }; destination messages       { file("/var/log/messages"); }; destination routers        { file("/var/log/routers.log"); }; destination secure         { file("/var/log/secure"); }; destination spooler        { file("/var/log/spooler"); }; destination syslog         { file("/var/log/syslog"); }; destination user           { file("/var/log/user.log"); }; #---------------------------------------------------------------------- #  Special catch all destination sorting by host #---------------------------------------------------------------------- destination hosts          { file("/var/log/HOSTS/$HOST/$YEAR/$MONTH/$DAY/$FACILITY_$HOST_$YEAR_$MONTH_$DAY"                  owner(root) group(root) perm(0600) dir_perm(0700) create_dirs(yes)); }; #---------------------------------------------------------------------- #  Forward to a loghost server #---------------------------------------------------------------------- #destination loghost       { udp("10.1.1.254" port(514)); }; #---------------------------------------------------------------------- #  Mail subsystem logs #---------------------------------------------------------------------- destination mail           { file("/var/log/mail.log"); }; destination mailerr        { file("/var/log/mail/errors"); }; destination mailinfo       { file("/var/log/mail/info"); }; destination mailwarn       { file("/var/log/mail/warnings"); }; #---------------------------------------------------------------------- #  INN news subsystem #---------------------------------------------------------------------- destination newscrit       { file("/var/log/news/critical"); }; destination newserr        { file("/var/log/news/errors"); }; destination newsnotice     { file("/var/log/news/notice"); }; destination newswarn       { file("/var/log/news/warnings"); }; #---------------------------------------------------------------------- #  Cron subsystem #---------------------------------------------------------------------- destination cron           { file("/var/log/cron.log"); }; destination crondebug      { file("/var/log/cron/debug"); }; destination cronerr        { file("/var/log/cron/errors"); }; destination croninfo       { file("/var/log/cron/info"); }; destination cronwarn       { file("/var/log/cron/warnings"); }; #---------------------------------------------------------------------- #  LPR subsystem #---------------------------------------------------------------------- destination lpr            { file("/var/log/lpr.log"); }; destination lprerr         { file("/var/log/lpr/errors"); }; destination lprinfo        { file("/var/log/lpr/info"); }; destination lprwarn        { file("/var/log/lpr/warnings"); }; #---------------------------------------------------------------------- #  Kernel messages #---------------------------------------------------------------------- destination kern           { file("/var/log/kern.log"); }; destination kernerr        { file("/var/log/kernel/errors"); }; destination kerninfo       { file("/var/log/kernel/info"); }; destination kernwarn       { file("/var/log/kernel/warnings"); }; #---------------------------------------------------------------------- #  Daemon messages #---------------------------------------------------------------------- destination daemon         { file("/var/log/daemon.log"); }; destination daemonerr      { file("/var/log/daemons/errors"); }; destination daemoninfo     { file("/var/log/daemons/info"); }; destination daemonwarn     { file("/var/log/daemons/warnings"); }; #---------------------------------------------------------------------- #  Console warnings #---------------------------------------------------------------------- destination console        { file("/dev/tty12"); }; #---------------------------------------------------------------------- #  All users #---------------------------------------------------------------------- destination users          { usertty("*"); }; #---------------------------------------------------------------------- #  Examples of programs that accept syslog messages and do something #  programatically with them. #---------------------------------------------------------------------- #destination mail-alert     { program("/usr/local/bin/syslog-mail"); }; #destination mail-perl      { program("/usr/local/bin/syslog-mail-perl"); }; #---------------------------------------------------------------------- #  Piping to Swatch #---------------------------------------------------------------------- #destination swatch         { program("/usr/bin/swatch --read-pipe=\"cat /dev/fd/0\""); }; #---------------------------------------------------------------------- #  Database notes: # #  Overall there seems to be three primary methods of putting data from #  syslog-ng into a database.  Each of these has certain pros and cons. # #  FIFO file:    Simply piping the template data into a First In, First #                Out file.  This will create a stream of data that will #                not require any sort of marker or identifier of how #                much data has been read.  This is the most elegant of #                the solutions and probably the most unstable. # #                Pros:  Very fast data writes and reads.  Data being #                       inserted into a database will be near real #                       time. # #                Cons:  Least stable of all the possible solutions, #                       and could require a lot of custom work to #                       make function on any particular Unix system. # #                       Loss of the pipe file will cause complete #                       data loss, and all following data that would #                       have been written to the FIFO file. # #  Buffer file:  While very similar to a FIFO file this is would be a #                text file which would buffer all the template #                output information.  Another program from cron or #                similar service would then run and source the buffer #                files and process the data into the database. # #                Pros:  Little chance of losing data since everything #                       will be written to a physical file much like #                       the regular logging process.   # #                       This method gives a tremendous amount of #                       flexibility since there would be yet another #                       opportunity to filter logs prior to inserting #                       any data into the database. # #                Cons:  Because there must be some interval between #                       the processing of the buffer files, there will #                       be a lag before the data is inserted in to the #                       database.   # #                       There is also a slight chance of data corruption #                       (ie bad insert command) if the system crashes #                       during a write, although this scenero is very #                       unlikely.   # #                       Another possible issue is that because multiple #                       buffer files be written, the previously run #                       sourcing file could get behind the data #                       insertion if there is a very large quantity of #                       logs being written.  This will totally depend #                       on the system that this is running on. # #  Program:      The least elegant of the solutions.  This method is to #                send the stream of data through some further interrupter #                program such as something in Perl or C.  That program #                will then take some action based off the data which #                could include writing to a database similarly to the #                program "sqlsyslogd". # #                Pros:  Allows complete control of the data, and as much #                       post processing as required. # #                Cons:  Slowest of all the forms.  Since the data will #                       have to go through some post processing it will #                       cause data being written to the database to #                       remain behind actual log records.  This could #                       cause a race condition in that logging is lost #                       either due to system crash, or high load on #                       the logging system. # #---------------------------------------------------------------------- #---------------------------------------------------------------------- #  Writing to a MySQL database: # #  Assumes a table/database structure of: # #            CREATE DATABASE syslog; #            USE syslog; # #            CREATE TABLE logs ( host varchar(32) default NULL, #                                facility varchar(10) default NULL, #                                priority varchar(10) default NULL, #                                level varchar(10) default NULL, #                                tag varchar(10) default NULL, #                                date date default NULL, #                                time time default NULL, #                                program varchar(15) default NULL, #                                msg text, seq int(10) unsigned NOT NULL auto_increment, #                                PRIMARY KEY (seq), #                                KEY host (host), #                                KEY seq (seq), #                                KEY program (program), #                                KEY time (time), #                                KEY date (date), #                                KEY priority (priority), #                                KEY facility (facility)) #                                TYPE=MyISAM; # #---------------------------------------------------------------------- #  Piping method #---------------------------------------------------------------------- #destination database       { pipe("/tmp/mysql.pipe" #                             template("INSERT INTO logs (host, facility, #                             priority, level, tag, date, time, program, #                             msg) VALUES ( '$HOST', '$FACILITY', '$PRIORITY', #                             '$LEVEL', '$TAG', '$YEAR-$MONTH-$DAY', #                             '$HOUR:$MIN:$SEC', '$PROGRAM', '$MSG' );\n") #                             template-escape(yes)); }; #---------------------------------------------------------------------- #  Buffer file method #---------------------------------------------------------------------- destination database       { file("/var/log/dblog/fulllog.$YEAR.$MONTH.$DAY.$HOUR.$MIN.$SEC"                              template("INSERT INTO logs (host, facility,                              priority, level, tag, date, time, program,                              msg) VALUES ( '$HOST', '$FACILITY', '$PRIORITY',                              '$LEVEL', '$TAG', '$YEAR-$MONTH-$DAY',                              '$HOUR:$MIN:$SEC', '$PROGRAM', '$MSG' );\n")                  owner(root) group(root) perm(0600)                  dir_perm(0700) create_dirs(yes)                              template-escape(yes)); }; #---------------------------------------------------------------------- #  Program method (alternate using sqlsyslogd): # #  Notes:  This is not a bad process, but lacks very much flexibility #          unless more changes are made to the source of sqlsyslogd. #          This is because sqlsyslogd assumes the data in a larger #          object style instead of breaking it down into smaller #          columnar pieces. #---------------------------------------------------------------------- #destination database       { program("/usr/local/sbin/sqlsyslogd -u #                             sqlsyslogd -t logs sqlsyslogs2 -p"); }; #---------------------------------------------------------------------- #  Since we probably will not be putting ALL of our logs in the database #  we better plan on capturing that data that we will be discarding for #  later review to insure we did not throw anything away we really #  should have captured. #---------------------------------------------------------------------- destination db_discard     { file("/var/log/discard.log"); }; #---------------------------------------------------------------------- #  Filters #---------------------------------------------------------------------- # #  Functions: # #  Name               Synopsis                        Description #  --------------  ------------------------------  -------------------- #  facility        facility(facility[,facility])    Match messages #                                                  having one of the #                                                  listed facility code. #  filter          Call another filter rule and #                  evaluate its value #  host            host(regexp)                    Match messages by #                                                  using a regular #                                                  expression against #                                                  the hostname field #                                                  of log messages. #  level/priority  level(pri[,pri1..pri2[,pri3]])  Match messages based #                                                  on priority. #  match           Tries to match a regular #                  expression to the message #                  itself. #  program         program(regexp)                 Match messages by #                                                  using a regular #                                                  expression against #                                                  the program name #                                                  field of log messages #---------------------------------------------------------------------- #  NOTES: # #  Getting filtering to work right can be difficult because while the #  syntax is fairly simple, it is not well documented.  To illustrate #  a brief lesson on filtering and to explain the majority of the #  mechanics, we shall use the filter from the PostgreSQL database #  how-to page found at:  http://www.umialumni.com/~ben/SYSLOG-DOC.html #                             #  This is a perfect and somewhat complex example to use.  In its #  original form it resembles: #                             #  filter f_postgres { not( #                            (host("syslogdb") and facility(cron) and level(info)) #                         or (facility(user) and level(notice) #                                 and ( match(" gethostbyaddr: ") #                                       or match("last message repeated ") #                                      ) #                             ) #                         or ( facility(local3) and level(notice) #                               and match(" SYSMON NORMAL ")) #                         or ( facility(mail) and level(warning) #                               and match(" writable directory") #                            ) #                         or (  ( host("dbserv1.somecompany.com") #                                 or host("dbserv2.somecompany.com") #                               ) #                               and facility(auth) and level(info) #                               and match("su oracle") and match(" succeeded for root on /dev/") #                            ) #                         ); }; # #  While in this form, it does not induce a tremendous amount of #  insight on what the specific filter is attempting to accomplish.  In #  reformatting the filter to resemble something a bit more human #  readable, it would look like: # #  filter f_postgres { not #                        ( #                          ( #                            host("syslogdb") and #                            facility(cron) and #                            level(info) #                          ) or #                          ( #                            facility(user) and #                            level(notice) and #                            ( #                              match(" gethostbyaddr: ") or #                              match("last message repeated ") #                            ) #                        ) or #                        ( #                          facility(local3) and #                          level(notice) and #                          match(" SYSMON NORMAL ") #                        ) or #                        ( #                          facility(mail) and #                          level(warning) and #                          match(" writable directory") #                        ) or #                        (   #                          ( #                            host("dbserv1.somecompany.com") or #                            host("dbserv2.somecompany.com") #                          ) and #                          facility(auth) and #                          level(info) and #                          match("su oracle") and #                          match(" succeeded for root on /dev/") #                        ) #                      ); #                   }; # #  Now in this form we can now begin to see what this filter has been #  attempting to accomplish.  We can now further breakdown each logical #  section and explain the different methods: # #  [1]  As in all statements in syslog-ng, each of the beginnings and #       endings must be with a curly bracket "{" "}" to clearly denote #       the start and finish. # #       In this filter, the entire filter is preferred by a "not" to #       indicate that these are the messages that we are NOT interested #       in and should be the ones filtered out.  All lines of logs that #       do not match these lines will be sent to the destination. # #  { not # #  [2]  The first major part of the filter is actually a compound #       filter that has two parts.  Because the two parts are separated #       by an "or", only one of the two parts must be matched for that #       line of log to be filtered. # #  [2a]  In the first part of this filter there are three requirements #        to be met for the filter to take affect.  These are the host #        string "syslogdb". the facility "cron", and the syslog level #        of info. # #  ( #    ( #      host("syslogdb") and #      facility(cron) and #      level(info) #    ) or # #  [2b]  In the second part of the filter, which in itself is a #        compound filter, there are three requirements as well.  These #        are that the facility of "user", and the log level of "notice" #        are met in addition to one of the two string matches that are #        shown in the example. # #  ( #    facility(user) and #    level(notice) and #    ( #      match(" gethostbyaddr: ") or #      match("last message repeated ") #    ) #  ) or # #  [3]  In the section of the filter there are once again three #       requirements to fire off a match which are a facility of "level3" #       a log level of "notice" and a sting match of " SYSMON NORMAL ". # #  ( #    facility(local3) and #    level(notice) and #    match(" SYSMON NORMAL ") #  ) or # #  [4]  This part of the filter is very similar to the previous #       filter, but with different search patterns. # #  ( #    facility(mail) and #    level(warning) and #    match(" writable directory") #  ) or # #  [5]  The last section of the filter is also a compound filter #       that to take affect will require that one of two hosts #       are matched, the facility of "auth", and log level of #       "info" occur in addition to the two string matches. # #  (   #    ( #      host("dbserv1.somecompany.com") or #      host("dbserv2.somecompany.com") #    ) and #    facility(auth) and #    level(info) and #    match("su oracle") and #    match(" succeeded for root on /dev/") #  ) # #  [6]  As in all command sets in syslog-ng, each of the statements #       must be properly closed with the correct ending punctuation #       AND a semi-colon.  Do not forget both, or you will be faced with #       an error. # #  ); }; # #  While this may not be the most complete example, it does cover the #  majority of the options and features that are available within the #  current version of syslog-ng. #---------------------------------------------------------------------- #---------------------------------------------------------------------- #  Standard filters for the standard destinations. #---------------------------------------------------------------------- filter      f_auth         { facility(auth, authpriv); }; filter      f_authpriv     { facility(authpriv); }; filter      f_cron         { facility(cron); }; filter      f_daemon       { facility(daemon); }; filter      f_kern         { facility(kern); }; filter      f_local1       { facility(local1); }; filter      f_local2       { facility(local2); }; filter      f_local3       { facility(local3); }; filter      f_local4       { facility(local4); }; filter      f_local5       { facility(local5); }; filter      f_local6       { facility(local6); }; filter      f_local7       { facility(local7); }; filter      f_lpr          { facility(lpr); }; filter      f_mail         { facility(mail); }; filter      f_messages     { facility(daemon, kern, user); }; filter      f_news         { facility(news); }; filter      f_spooler      { facility(uucp,news) and level(crit); }; filter      f_syslog       { not facility(auth, authpriv) and not facility(mail); }; filter      f_user         { facility(user); }; #---------------------------------------------------------------------- #  Other catch-all filters #---------------------------------------------------------------------- filter      f_crit         { level(crit); }; #filter     f_debug        { not facility(auth, authpriv, news, mail); }; filter      f_debug        { level(debug); }; filter      f_emergency    { level(emerg); }; filter      f_err          { level(err); }; filter      f_info         { level(info); }; filter      f_notice       { level(notice); }; filter      f_warn         { level(warn); }; #---------------------------------------------------------------------- #  Filer for the MySQL database pipe.  These are things that we really #  do not care to see otherwise they may fill up our database with #  garbage. #---------------------------------------------------------------------- #filter      f_db           { not facility(kern) and level(info, warning) or #                             not facility(user) and level(notice) or #                             not facility(local2) and level(debug); }; # #filter      f_db           { not match("last message repeated ") or #                             not match("emulate rawmode for keycode"); }; # #filter      f_discard      { facility(kern) and level(info, warning) or #                             facility(user) and level(notice) or #                 facility(local2) and level(debug); }; # #filter      f_discard      { match("last message repeated ") or #                             match("emulate rawmode for keycode"); }; #---------------------------------------------------------------------- #  Logging #---------------------------------------------------------------------- # #  Notes:  When applying filters, remember that each subsequent filter #          acts as a filter on the previous data flow.  This means that #          if the first filter limits the flow to only data from the #          auth system, a subsequent filter for authpriv will cause #          no data to be written.  An example of this would be: # # log { source(s_dgram); #       source(s_internal); #       source(s_kernel); #       source(s_tcp); #       source(s_udp);      filter(f_auth); #                           filter(f_authpriv);  destination(authlog); }; # #          So, one can cancel out the other. # #  There are also certain flags that can be attached to each of the log #  statements: # #  Flag      Description #  --------  ---------------------------------------------------------- #  catchall  This flag means that the source of the message is ignored, #            only the filters are taken into account when matching #            messages. #  fallback  This flag makes a log statement 'fallback'. Being a #            fallback statement means that only messages not matching #            any 'non-fallback' log statements will be dispatched. #  final     This flag means that the processing of log statements ends #            here. Note that this doesn't necessarily mean that #            matching messages will be stored once, as they can be #            matching log statements processed prior the current one. #---------------------------------------------------------------------- #---------------------------------------------------------------------- #  Standard logging #---------------------------------------------------------------------- log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_auth);      destination(authlog); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_local7);    destination(bootlog); }; #log{ source(s_dgram); #      source(s_internal); #      source(s_kernel); #      source(s_tcp); #      source(s_udp);      filter(f_debug);     destination(debug); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_local1);    destination(explan); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_local5);    destination(routers); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_messages);  destination(messages); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_authpriv);  destination(secure); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_spooler);   destination(spooler); }; log { source(s_dgram);       source(s_internal);       source(s_kernel);       source(s_tcp);      filter(f_syslog);    destination(syslog); }; #log { source(s_dgram); #      source(s_internal); #      source(s_kernel); #      source(s_tcp); #      source(s_udp);                       destination(syslog); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_user);      destination(user); }; #---------------------------------------------------------------------- #  Special catch all destination sorting by host #---------------------------------------------------------------------- log { source(s_dgram);       source(s_internal);       source(s_kernel);       source(s_tcp);                           destination(hosts); }; #---------------------------------------------------------------------- #  Send to a loghost #---------------------------------------------------------------------- #log { source(s_dgram); #      source(s_internal); #      source(s_kernel); #      source(s_tcp);                           destination(loghost); }; #---------------------------------------------------------------------- #  Mail subsystem logging #---------------------------------------------------------------------- #log { source(s_dgram); #      source(s_internal); #      source(s_kernel); #      source(s_tcp); #      source(s_udp);     filter(f_mail);      destination(mail); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_mail);                           filter(f_err);       destination(mailerr); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_mail);                           filter(f_info);      destination(mailinfo); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_mail);                           filter(f_notice);    destination(mailinfo); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_mail);                           filter(f_warn);      destination(mailwarn); }; #---------------------------------------------------------------------- #  INN subsystem logging #---------------------------------------------------------------------- log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_news);                           filter(f_crit);      destination(newscrit); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_news);                           filter(f_err);       destination(newserr); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_news);                           filter(f_notice);    destination(newsnotice); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_news);                           filter(f_warn);      destination(newswarn); }; #---------------------------------------------------------------------- #  Cron subsystem logging #---------------------------------------------------------------------- #log { source(s_dgram); #      source(s_internal); #      source(s_tcp); #      source(s_udp);     filter(f_cron);      destination(crondebug); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_cron);                           filter(f_err);       destination(cronerr); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_cron);                           filter(f_info);      destination(croninfo); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_cron);                           filter(f_warn);      destination(cronwarn); }; #---------------------------------------------------------------------- #  LPR subsystem logging #---------------------------------------------------------------------- #log { source(s_dgram); #      source(s_internal); #      source(s_tcp); #      source(s_udp);     filter(f_lpr);       destination(lpr); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_lpr);                           filter(f_err);       destination(lprerr); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_lpr);                           filter(f_info);      destination(lprinfo); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_lpr);                              filter(f_warn);      destination(lprwarn); }; #---------------------------------------------------------------------- #  Kernel subsystem logging #---------------------------------------------------------------------- #log { source(s_dgram); #      source(s_internal); #      source(s_kernel); #      source(s_tcp); #      source(s_udp);     filter(f_kern);      destination(kern); }; log { source(s_dgram);       source(s_internal);       source(s_kernel);       source(s_tcp);      filter(f_kern);                           filter(f_err);       destination(kernerr); }; log { source(s_dgram);       source(s_internal);       source(s_kernel);       source(s_tcp);      filter(f_kern);                           filter(f_info);      destination(kerninfo); }; log { source(s_dgram);       source(s_internal);       source(s_kernel);       source(s_tcp);      filter(f_kern);                           filter(f_warn);      destination(kernwarn); }; #---------------------------------------------------------------------- #  Daemon subsystem logging #---------------------------------------------------------------------- #log { source(s_dgram); #      source(s_internal); #      source(s_tcp); #      source(s_udp);     filter(f_daemon);    destination(daemon); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_daemon);                               filter(f_err);       destination(daemonerr); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_daemon);                           filter(f_info);      destination(daemoninfo); }; log { source(s_dgram);       source(s_internal);       source(s_tcp);      filter(f_daemon);                               filter(f_warn);      destination(daemonwarn); }; #---------------------------------------------------------------------- #  Console logging #---------------------------------------------------------------------- #  16-Mar-03 - REP - Removed logging to the console for performance #                    reasons.  Since we are not really going to be #                    looking at the console all the time, why log there #                    anyway. #---------------------------------------------------------------------- #log { source(s_dgram); #      source(s_internal); #      source(s_kernel); #      source(s_tcp);      filter(f_syslog);    destination(console); }; #---------------------------------------------------------------------- #  Logging to a database #---------------------------------------------------------------------- #log { source(s_dgram); #      source(s_internal); #      source(s_kernel); #      source(s_tcp);      filter(f_db);        destination(database); }; #log { source(s_dgram); #      source(s_internal); #      source(s_kernel); #      source(s_tcp);      filter(f_discard);   destination(db_discard); }; /etc/init.d/syslog-ng start