- syslog-ng - lists.balabit.hu

(U) Find the source systems that write to specific Syslog Server
by Amin, Jitesh CTR DISA JSP (US) 03 Jul '18

03 Jul '18

CLASSIFICATION: UNCLASSIFIED Hello, We have multiple servers running syslog. By looking at the syslong.conf file we can identify where the syslog servers are forwarding the data to. But what we really want to know is what all sources are writing their logs to our syslog servers. Is there a way for us to look somewhere within syslog configuration and find out which all systems are forwarding/writing logs to a specific syslog server? Thanks Jitesh AminCLASSIFICATION: UNCLASSIFIED

2 1

Re: [syslog-ng] syslog-ng not following symlinks correctly on UBUNTU, works fine in RHEL
by Donatello D 02 Jul '18

02 Jul '18

@Jim - this is what i use as a workaround already, but it is sub-optimal, as there will be other files that are rolled over in different intervals, so i end up reloading config multiple times. the real question is why does it work in RHEL and fail in UBUNTU? > Date: Sat, 30 Jun 2018 20:31:59 -0400 > From: Jim Hendrick <james.r.hendrick(a)gmail.com> > To: "Syslog-ng users' and developers' mailing list" > <syslog-ng(a)lists.balabit.hu> > Subject: Re: [syslog-ng] syslog-ng not following symlinks correctly on > UBUNTU, works fine in RHEL > Message-ID: > <CANEn2idABV25G1vFa4B=WhOyuHjd3HwLMKFBHgqydH6zvH0H9w(a)mail.gmail.com> > Content-Type: text/plain; charset="utf-8" > > hardlinks are simply additional pointers to the same inode. > symlinks contain the name of the referenced object > > If you want a hack - maybe have syslog-ng reload after the file changes? > (SIGHUP) > > > > On Sat, Jun 30, 2018 at 12:56 PM, Ankit Agarwal <ankit(a)travelmyheart.org> > wrote: > > > Hi, > > > > I ran into a similar problem on Ubuntu as well. > > > > In my case, I was tracking the Tomcat localhost log file in syslog-ng but > > Tomcat creates a new log file everyday by default, and the filename changes > > (since it includes the date). > > > > Therefore, I periodically created a softlink to the localhost log file > > where the link had a constant name. The constant name is needed because I > > obviously cannot keep changing the syslog-ng configuration to match the > > day's localhost log file name. > > > > I found that the softlink did not work. > > > > Instead I had to create a hardlink. > > > > This is because the softlink's modified date does not change when the > > underlying file changes. The hardlink's modified date does change since it > > is pointing to the actual data. We need the modified date to change for the > > syslog-ng client to pick up new log entries. > > > > In my case, I periodically ran the following command via CRON in the > > Tomcat logs directory: > > > > sudo ln -f $(ls -t localhost.* | head -1) tomcat_localhost.log > > > > This is to get the latest localhost log file and create the hardlink for > > it (overwriting the older hardlink that may have been pointing to the > > previous day's localhost log file). > > > > I ran this every hour just to be safe. > > > > > > So in your case, I think you would just need to recreate the hardlink as > > soon as your log file is rotated. > > > > > > Hope this helps. > > > > Ankit > > > > > > > > ---- On Sat, 30 Jun 2018 01:13:44 -0700 *Donatello D > > <bluray.vik(a)gmail.com <bluray.vik(a)gmail.com>>* wrote ---- > > > > syslog-ng is configured to read a symlink pointing to logs generated from > > my application which rotates the file using log4j2 rollingfile appender. > > Everything works fine till the rotation happens. after the file get rotated > > syslog-ng still seems to hold on to the older inode (which is not moved) > > and doesn't change to follow the new logs. this however does not happen in > > RHEL where syslog-ng recognizes the file is now rotated and moves to the > > new file. In both cases the sym link is always configured to point to the > > latest file. version details and logs from both OSs below. > > > > What am i missing here? > > > > UBUNTU - > > syslog-ng 3.5.6 > > Installer-Version: 3.5.6 > > Revision: 3.5.6-2.1 [@416d315] (Ubuntu/16.04) > > Compile-Date: Oct 24 2015 03:49:19 > > Available-Modules: afsocket,afuser,tfgeoip,confgen,csvparser, > > syslogformat,afamqp,redis,afsql,affile,afsmtp,linux- > > kmsg-format,dbparser,system-source,cryptofuncs,basicfuncs, > > json-plugin,afprog,afsocket-tls,afstomp,afsocket-notls,afmongodb > > Enable-Debug: off > > Enable-GProf: off > > Enable-Memtrace: off > > Enable-IPv6: on > > Enable-Spoof-Source: on > > Enable-TCP-Wrapper: on > > Enable-Linux-Caps: on > > Enable-Pcre: on > > > > symlink is pointing to the file that gets the logs. prior to rotation the > > process watches correctly for the file (same inodes held by my app and > > syslog-ng) > > > > lrwxrwxrwx 1 root root 56 Jun 29 08:44 node1-access.log -> > > /x/logs/vik-test_access.log > > > > COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME > > java 11032 vikram 53w REG 8,1 1101 1542626 > > vik-test_access.log > > syslog-ng 21661 root 9r REG 8,1 1101 1542626 > > vik-test_access.log > > > > > > Post rotation, syslog-ng holds on to the older file (now rotated). > > > > COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME > > java 11032 vikram 53w REG 8,1 876 1542631 > > e/elasticsearch-6.2.3/logs/vik-test_access.log > > syslog-ng 21661 root 9r REG 8,1 1101 1542626 > > e/elasticsearch-6.2.3/logs/vik-test_access-2018-06-30.log > > > > The same setup works perfectly fine in RHEL (version details below) where > > syslog-ng follows the new file correctly. > > > > RHEL > > syslog-ng 3.3.5 > > Installer-Version: 3.3.5 > > Revision: ssh+git://bazsi@git.balabit//var/scm/git/syslog-ng/syslog- > > ng-ose--mainline--3.3--master#d5d607c05251b38e821efe27bc46ac8db78dd722 > > Compile-Date: Oct 18 2012 15:17:09 > > Default-Modules: affile,afprog,afsocket,afuser, > > basicfuncs,csvparser,dbparser,syslogformat > > Available-Modules: afprog,afsocket-tls,dbparser,confgen,convertfuncs, > > basicfuncs,afsocket,afmongodb,csvparser,affile,dummy,syslogformat,afuser > > Enable-Debug: off > > Enable-GProf: off > > Enable-Memtrace: off > > Enable-IPv6: on > > Enable-Spoof-Source: off > > Enable-TCP-Wrapper: on > > Enable-Linux-Caps: off > > Enable-Pcre: on > > > > ____________________________________________________________ > > __________________ > > Member info: https://lists.balabit.hu/mailman/listinfo/syslog-ng > > Documentation: http://www.balabit.com/support/documentation/? > > product=syslog-ng > > FAQ: http://www.balabit.com/wiki/syslog-ng-faq > > > > > > > > > > ____________________________________________________________ > > __________________ > > Member info: https://lists.balabit.hu/mailman/listinfo/syslog-ng > > Documentation: http://www.balabit.com/support/documentation/? > > product=syslog-ng > > FAQ: http://www.balabit.com/wiki/syslog-ng-faq > > > > > > >

4 6

syslog-ng not following symlinks correctly on UBUNTU, works fine in RHEL
by Donatello D 01 Jul '18

01 Jul '18

syslog-ng is configured to read a symlink pointing to logs generated from my application which rotates the file using log4j2 rollingfile appender. Everything works fine till the rotation happens. after the file get rotated syslog-ng still seems to hold on to the older inode (which is not moved) and doesn't change to follow the new logs. this however does not happen in RHEL where syslog-ng recognizes the file is now rotated and moves to the new file. In both cases the sym link is always configured to point to the latest file. version details and logs from both OSs below. What am i missing here? UBUNTU - syslog-ng 3.5.6 Installer-Version: 3.5.6 Revision: 3.5.6-2.1 [@416d315] (Ubuntu/16.04) Compile-Date: Oct 24 2015 03:49:19 Available-Modules: afsocket,afuser,tfgeoip,confgen,csvparser,syslogformat,afamqp,redis,afsql,affile,afsmtp,linux-kmsg-format,dbparser,system-source,cryptofuncs,basicfuncs,json-plugin,afprog,afsocket-tls,afstomp,afsocket-notls,afmongodb Enable-Debug: off Enable-GProf: off Enable-Memtrace: off Enable-IPv6: on Enable-Spoof-Source: on Enable-TCP-Wrapper: on Enable-Linux-Caps: on Enable-Pcre: on symlink is pointing to the file that gets the logs. prior to rotation the process watches correctly for the file (same inodes held by my app and syslog-ng) lrwxrwxrwx 1 root root 56 Jun 29 08:44 node1-access.log -> /x/logs/vik-test_access.log COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME java 11032 vikram 53w REG 8,1 1101 1542626 vik-test_access.log syslog-ng 21661 root 9r REG 8,1 1101 1542626 vik-test_access.log Post rotation, syslog-ng holds on to the older file (now rotated). COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME java 11032 vikram 53w REG 8,1 876 1542631 e/elasticsearch-6.2.3/logs/vik-test_access.log syslog-ng 21661 root 9r REG 8,1 1101 1542626 e/elasticsearch-6.2.3/logs/vik-test_access-2018-06-30.log The same setup works perfectly fine in RHEL (version details below) where syslog-ng follows the new file correctly. RHEL syslog-ng 3.3.5 Installer-Version: 3.3.5 Revision: ssh+git://bazsi@git.balabit //var/scm/git/syslog-ng/syslog-ng-ose--mainline--3.3--master#d5d607c05251b38e821efe27bc46ac8db78dd722 Compile-Date: Oct 18 2012 15:17:09 Default-Modules: affile,afprog,afsocket,afuser,basicfuncs,csvparser,dbparser,syslogformat Available-Modules: afprog,afsocket-tls,dbparser,confgen,convertfuncs,basicfuncs,afsocket,afmongodb,csvparser,affile,dummy,syslogformat,afuser Enable-Debug: off Enable-GProf: off Enable-Memtrace: off Enable-IPv6: on Enable-Spoof-Source: off Enable-TCP-Wrapper: on Enable-Linux-Caps: off Enable-Pcre: on

3 2

Random crash in the main thread
by som shukla 28 Jun '18

28 Jun '18

Hi Guys, Recently we came up with a crash in syslog-ng with following back trace. rogram terminated with signal SIGBUS, Bus error. #0 0x00007f03e00b4f7c in ?? () from /usr/lib64/libsyslog-ng.so.0 (gdb) (gdb) (gdb) (gdb) (gdb) #0 0x00007f03e00b4f7c in ?? () from /usr/lib64/libsyslog-ng.so.0 No symbol table info available. #1 0x00007f03e00b5a27 in persist_state_start () from /usr/lib64/libsyslog-ng.so.0 No symbol table info available. #2 0x00007f03e00a3d7a in cfg_initial_init () from /usr/lib64/libsyslog-ng.so.0 No symbol table info available. #3 0x00000000004028e0 in initial_init () No symbol table info available. #4 0x0000000000402a36 in main () I thought it to be similar to the this issue https://github.com/balabit/syslog-ng/issues/716 ,but I'm not sure if this is really the same. Has somebody come up with similar issue Thanks Somji

2 3

CentOS 7.5, rsyslog, port 601, TCP - nothing getting delivered
by Lachlan Musicman 26 Jun '18

26 Jun '18

I don't understand why this isn't working? I'm not seeing any data in our Balabit appliance. I have a regular default installation of CentOS 7.5, and have followed the RedHat 7 rsyslog directions with regard to setting up a new message filter: I've added a singe file to /etc/rsyslog.d/ [root@host02 /etc/rsyslog.d]# cat tcp601.conf *.* action(type="omfwd" queue.type="LinkedList" queue.filename="example_fwd_tcp_601" action.resumeRetryCount="-1" queue.saveonshutdown="on" template="RSYSLOG_SyslogProtocol23Format" target="10.126.19.45" Port="601" Protocol="tcp") But I'm not getting anything at the appliance? The Appliance Log Source seems to be set up correctly (no licensing issues, port 601 is set, Syslog format (I was told that is RFC 5425) selected). Ports are open, but on the server that's configured as per above, I'm seeing this: [root@host02 log]# netstat -tnp| grep 601 tcp 1 0 10.126.19.66:39768 10.126.19.45:601 CLOSE_WAIT 2400/rsyslogd The data works fine if I send over UDP/port 514, with the template being either RSYSLOG_SyslogProtocol23Format or RSYSLOG_TraditionalFileFormat I'm also seeing - in host02's /var/log/messages a *large* number of errors that state: Jun 25 11:14:14 host02 rsyslogd: action 'action 2' resumed (module 'builtin:omfwd') [v8.24.0 try http://www.rsyslog.com/e/2359 ] Can someone tell me where I've gone wrong and/or indicate what I might do next to debug this issue? Cheers L.

2 3

syslog-ng parsing Error
by David Campeau 25 Jun '18

25 Jun '18

Hello, I have a syslog source node sending syslogs, and they are being generated via a python script, and is using Python Rfc5426SysLogHandler. So, these syslog messages should be RFC compliant. However, syslog-ng does prepend an error message before sending it on to be put into storage. Example error message from syslog-ng = <43>Jun 21 10:27:38 syslog-ng-Server syslog-ng[2559]: Error processing log message: xxxxx timestamp, source hostname and payload follows. I've done some googling, but haven't been able to find out what error 2559 means. Any thoughts of what I might do to determine what syslog-ng isn't liking about the syslog it is receiving? I need to relay this information to a developer so they can make adjustments to the python script. Best regards, David

2 4

getting error destination plugin java not found while using mod-java
by Pratik Sharma 25 Jun '18

25 Jun '18

Hi Team, I have a setup of syslog-ng in Centos 7 with jdk 1.8.To use a custom java program i have installed syslog-ng-java plugin using yum install. I have added a custom syslog-ng file in conf.d directory.On starting the syslog-ng service i am reciving error destination java plugin not found. *ERROR:* > Jun 24 14:16:01 localhost syslog-ng: Error parsing destination, > destination plugin java not found in > /etc/syslog-ng/conf.d/syslog-ng-custom.conf at line 15, column 3: > Jun 24 14:16:01 localhost syslog-ng: included from > /etc/syslog-ng/syslog-ng.conf line 68, column 1 > Jun 24 14:16:01 localhost syslog-ng: java( > Jun 24 14:16:01 localhost syslog-ng: ^^^^ > Syslog-ng is detecting the java plugin(verified by syslog-ng -V) *syslog-ng -V:* > [root@localhost server]# syslog-ng -V >>>>> syslog-ng 3 (3.14.1) >>>>> Config version: 3.14 >>>>> Installer-Version: 3.14.1 >>>>> Revision: >>>>> Compile-Date: Feb 28 2018 07:34:20 >>>>> Module-Directory: //usr/lib64/syslog-ng >>>>> Module-Path: //usr/lib64/syslog-ng >>>>> Available-Modules: >>>>> add-contextual-data,afamqp,affile,afprog,afsocket,afstomp,afuser,appmodel,basicfuncs,cef,confgen,cryptofuncs,csvparser,date,dbparser,disk-buffer,graphite,json-plugin,kvformat,linux-kmsg-format,map-value-pairs,pseudofile,sdjournal,snmptrapd-parser,stardate,syslogformat,system-source,tags-parser,tfgetent,xml, >>>>> *mod-java* >>>>> Enable-Debug: off >>>>> Enable-GProf: off >>>>> Enable-Memtrace: off >>>>> Enable-IPv6: on >>>>> Enable-Spoof-Source: on >>>>> Enable-TCP-Wrapper: on >>>>> Enable-Linux-Caps: on >>>>> Enable-Systemd: on >>>>> >>>> Java Version: > [root@localhost server]# java -version > java version "1.8.0_171" > Java(TM) SE Runtime Environment (build 1.8.0_171-b11) > Java HotSpot(TM) 64-Bit Server VM (build 25.171-b11, mixed mode) > *conf.d/syslog-ng-custom.conf* source s_apache { > system(); # Check which OS & collect system logs > internal(); > }; > > destination d_remote { > tcp("1.2.3.4" port(514)); > }; > destination d_device_file { > file("/var/log/$HOST.$MIN.log"); > }; > > destination java_to_file{ > java( > class_name("SampleJavaDestination") > class_path("/root") > option("name", "value") > option("filepath", "/var/log/example.txt") > ); > }; > > log { > source(s_apache); > destination(d_device_file); > destination(java_to_file); > > }; > Can someone please tell me what i am missing?

3 2

Fwd: SIGBUS error patch version
by som shukla 22 Jun '18

22 Jun '18

---------- Forwarded message ---------- From: som shukla <somj.shukla(a)gmail.com> Date: Fri, Jun 22, 2018 at 12:04 PM Subject: SIGBUS error patch version To: syslog-ng(a)lists.balabit.hu Hi , I'm using OSE version of the syslog-ng. There is a bug with persist file which has been solved recently in Feb. Following is the link for the githubpage https://github.com/balabit/syslog-ng/pull/1785/files. Was it ever included as patch in the standard versions. If yes then what is the version number. Thanks Somji

2 1

Poll: how quickly you update syslog-ng
by Czanik, Péter 20 Jun '18

20 Jun '18

Hi, I just posted a syslog-ng update related poll on Twitter: https://twitter.com/PCzanik/status/1009429698353553408 The reason is, that I'd like to learn how quickly you start to use a new version of syslog-ng after release. I see some rough download statistics for my Fedora / CentOS packages, but that is only a slice of syslog-ng users. Please answer the poll, and If you have any comments, alternative answers, explanations, you can leave it here or on Twitter. Also: retweet to reach more syslog-ng users! Thanks, Peter Czanik (CzP) <peter.czanik(a)balabit.com> Balabit / syslog-ng upstream https://syslog-ng.com/blog/author/peterczanik/ https://twitter.com/PCzanik

1 0

Re: [syslog-ng] hostname not appearing correctly when receiving logs from switches
by Clayton Dukes 19 Jun '18

19 Jun '18

Was anyone able to figure out why this is happening? From: "Scheidler, Balázs" <balazs.scheidler(a)balabit.com> Date: Tuesday, May 15, 2018 at 10:58 AM To: "cdukes(a)cdukes.com" <cdukes(a)cdukes.com> Cc: Clayton Dukes <cdukes(a)logzilla.net> Subject: Re: Fwd: [syslog-ng] hostname not appearing correctly when receiving logs from switches Thanks On May 15, 2018 16:52, "Clayton Dukes" <cdukes(a)gmail.com<mailto:cdukes@gmail.com>> wrote: Making sure you got this (in case the other went to spam) ______________________________________________________________ Clayton Dukes ______________________________________________________________ ---------- Forwarded message ---------- From: Clayton Dukes <cdukes(a)logzilla.net<mailto:cdukes@logzilla.net>> Date: Mon, May 7, 2018 at 4:46 PM Subject: Re: [syslog-ng] hostname not appearing correctly when receiving logs from switches To: "Scheidler, Balázs" <balazs.scheidler(a)balabit.com<mailto:balazs.scheidler@balabit.com>> Cc: Syslog-ng users' and developers' mailing list <syslog-ng(a)lists.balabit.hu<mailto:syslog-ng@lists.balabit.hu>> Hi Balazs, Sorry for the delay, I don’t get a lot of free time these days :) I have attached a pcap as well as a raw log. The log is prior to any manipulation of LogZilla rules, etc. This is easily reproduceable. Also, if I add show-timezone to the device config, the host field shows up. The problem, of course, is that we can’t tell all of our customers to re-configure all of their cisco devices. We have documented the work-around here (search the page for “hostname missing”): http://demo.logzilla.net/help/receiving_data/cisco_ios_configuration This problem never existed before, but I am not certain which syslog-ng version it started occurring in. From: "Scheidler, Balázs" <balazs.scheidler(a)balabit.com<mailto:balazs.scheidler@balabit.com>> Date: Tuesday, May 1, 2018 at 8:45 PM To: Clayton Dukes <cdukes(a)logzilla.net<mailto:cdukes@logzilla.net>> Cc: Joshua <aces621(a)yahoo.com<mailto:aces621@yahoo.com>>, Syslog-ng users' and developers' mailing list <syslog-ng(a)lists.balabit.hu<mailto:syslog-ng@lists.balabit.hu>> Subject: Re: [syslog-ng] hostname not appearing correctly when receiving logs from switches Interesting that I saw this message the first time in your response, and not the original one. Anyhow, to understand the problem we would need an exact byte-by-byte representation of what syslog-ng is receiving from the switch together with the configuration that is used to process it. A tcpdump or an "Incoming message" from syslog debug outout should work. We haven't intentionally changed the syslog parser as far as I remember. On May 1, 2018 22:50, "Clayton Dukes" <cdukes(a)logzilla.net<mailto:cdukes@logzilla.net>> wrote: Interesting! We’ve been getting a lot of support tickets for this very problem. I can easily recreate the issue. Balabit Team: is this a new bug? [cid:image001.png@01D306E3.0FEBC990] Clayton Dukes Founder & CEO LogZilla Corporation 2900 N. Quinlan Park Rd<https://maps.google.com/?q=2900+N.+Quinlan+Park+Rd&entry=gmail&source=g>, B240-341 Austin, TX, 78732 Tel: 936-4NetOps (463-8677) Web: www.logzilla.net<http://www.logzilla.net/> [cid:image002.png@01D306E3.0FEBC990]<https://twitter.com/logzilla>[cid:image003.png@01D306E3.0FEBC990]<https://youtu.be/drg5wv_mgfA>[cid:image004.png@01D306E3.0FEBC990]<https://www.linkedin.com/in/lzcdukes/> For NetOps, By NetOps! From: syslog-ng <syslog-ng-bounces(a)lists.balabit.hu<mailto:syslog-ng-bounces@lists.balabit.hu>> on behalf of Joshua <aces621(a)yahoo.com<mailto:aces621@yahoo.com>> Reply-To: Joshua <aces621(a)yahoo.com<mailto:aces621@yahoo.com>>, Syslog-ng users' and developers' mailing list <syslog-ng(a)lists.balabit.hu<mailto:syslog-ng@lists.balabit.hu>> Date: Monday, April 30, 2018 at 7:09 PM To: "syslog-ng(a)lists.balabit.hu<mailto:syslog-ng@lists.balabit.hu>" <syslog-ng(a)lists.balabit.hu<mailto:syslog-ng@lists.balabit.hu>> Subject: [Suspected Spam] [syslog-ng] hostname not appearing correctly when receiving logs from switches Hi All, I am pretty new to syslog-ng but do have some basic knowledge. I have deployed syslog-ng v3.14 on a newly deployed Linux server because syslog-ng v3.5 is working very well on another syslog server. On this new deployment, the syslogs received from most of the servers are able to show IP/host, however, the syslogs from our switches contains IP/host showing as ":" (colons). I copied the current working custom build .conf from another syslog server into our new server. Can someone help me figure out what I am missing? It is working for some components but not for switches. I tested the same switch by sending syslog to another syslog server and the hostname is appearing but just not appearing on the new syslog server. The only difference between the two server is that one uses v3.5 (the working one) and the other uses syslog-ng v3.14. I have set: "keep_hostname (yes)" but it still doesn't work. Can someone please help? Am I missing something here? Thanks Joshua Lai ______________________________________________________________________________ Member info: https://lists.balabit.hu/mailman/listinfo/syslog-ng Documentation: http://www.balabit.com/support/documentation/?product=syslog-ng FAQ: http://www.balabit.com/wiki/syslog-ng-faq

2 3