- syslog-ng - lists.balabit.hu

question about syslog-ng to MySQL
by Martin Hess 08 Jul '02

08 Jul '02

Hi all! I'm pretty new here in the hole syslog-ng world; so I have a litte question witch might sounds stupid for you... but any help would be great! I successfuly set up syslog-ng to log into the MySQL database. But now I want to filter the events before they are piped into the database (I don't want the logs from the cron jobs to be logged). Does anybody can tell me how I can achieve this?? Thanks in advance martin, switzerland

2 1

How do I stop "STATS: dropped 0" messages?
by Subodh Nijsure 05 Jul '02

05 Jul '02

Hello, I think I have read the man pages, and I missed how do I stop syslog-ng from generating the "STATS: dropped 0" messages? Is there option in config file I can specify? /Subodh

3 2

pipe() under hpux-11 and timestamp
by Trapp, Michael 05 Jul '02

05 Jul '02

hi, i am running syslog-ng 1.5.18 on hpux 11.00 with the following options: use_time_recvd(yes); use_fqdn(yes); sync(0); use_dns(yes); dns_cache(yes); dns_cache_size(10000); dns_cache_expire(3600); dns_cache_expire_failed(300); keep_hostname(no); chain_hostnames(no); gc_idle_threshold(50); gc_busy_threshold(5000); log_fifo_size(2000); reading from udp() and pipe() as far as i can see the pipe("/dev/log") isn't working correct. with a simple fgets() i can see that the messages are '0' terminated strings, but syslog-ng reads the whole message obviously each message has 4096 bytes and print out several lines filled up with characters from the buffer behind the string termination. i have also configured use_time_recvd(yes) but there are a few messages which don't have the right timestamp or are not in the right order (just a few seconds). in adition the day field of this lines has a leading '0'. has anybody experienced these problems? are there further options to improve the performance, as i can see syslog-ng still miss some packets. thanx michael

2 2

starting syslog-ng
by bert 04 Jul '02

04 Jul '02

Hi, I want to start syslog-ng before the network is running but my configuration script contains (among other destinations) a network reference. Hence syslog-ng is exiting due to an error in the initialisation. Is there a way to deal with this, so syslog-ng is starting and resolve the connection at the time the network starts up. Thanks, Bert

2 1

Long messages truncated in 1.5.x branch on Solaris 8
by Andreas Schulze 03 Jul '02

03 Jul '02

Hello, we are observing a wondrous behavior on our central logservers. We have an local application, that logs to the destinations above. And all messages greater that approx. 975 Bytes are truncated. But the enigma is: If using the local Solaris 'logger' application, all messages greater than 509 Bytes are truncated. But they are truncated on different positions depends on the message size! (Say, if you log 600 Bytes, you get the first 509 and if you try to log 761, you get the first 26.) Hmm ?! We tried out the log_msg_size() feature, available since 1.5.1?. With no luck. Both logservers using Solaris 8, one running syslog-ng 1.5.7, the other 1.5.18. We are using file() and pipe() destinations capable of queueing via log_fifo_size() and capable of formating output via template(..$MSG..\n). Before writing to the destinations, we do a lot of filtering, too. We are using the excellent syslog-ng a long time, but now we stuck inactive. Has anybody a hint for us? Is this a bug in syslog-ng, in Solaris, in our mind? Thanks a lot for any note. -- Best regards --Andreas Schulze [phone: +49.5246.80.1275, fax: +49.5246.80.2275] | I believe, it was Dennis Ritchie who said something like: | "C is rarely the best language for a given task, | but it's often the second-best". | The implication being that: "[...]" | | sh# cat>$$.c<<EOT | main(l,a,n,d)char**a;{for(d=atoi(a[1])/10*80-atoi(a[2])/5-596;n="@NK\ | ACLCCGZAAQBEAADAFaISADJABBA^SNLGAQABDAXIMBAACTBATAHDBANZcEMMCCCCAAhE\ | IJFAEAAABAfHJETBdFLDAANEfDNBPHdBcBBBEA_AL H E L L O, W O R L D! " | [l++-3];)for(;n-->64;)putchar(!d+++33^l&1);} | EOT | gcc -o$$ $$.c;clear;./$$ 52 8;rm -f $$*

2 5

Re: [syslog-ng]host x.x.x.111 is the same as x.x.x.11 (filter error?)
by Lubos Vitek 03 Jul '02

03 Jul '02

Thank you very much for your kind reply! It really works as you wrote :-) (with ^...$ RE). Now I have one another question. How can I filter out this kind of log record: datum hostX process some text with "xxxxx" string I want to write it in file for hostX.log, so I wrote filter: filter f_hostX { host("hostX"); }; then I wrote filter for string "xxxxx": filter f_notxxxxx { not match("xxxxx"); }; and then I used it all in log statement: log {... filter f_hostX ; filter f_notxxxxx ; ... } I tried to log something on host hostX with string "xxxxx", and I received it in syslog-ng in file hostX.log. Why ? What am I doing wrong now ? Please sorry if I missunderstand something, but I tried it more times with other combinations of filter, but the "NOT" statement simply does not work ... Thank you in advance for your reply... Bye, Lubos. __________________________________________________ Do You Yahoo!? Sign up for SBC Yahoo! Dial - First Month Free http://sbc.yahoo.com

2 1

host x.x.x.111 is the same as x.x.x.11 (filter error?)
by Lubos Vitek 02 Jul '02

02 Jul '02

Hallo, I have found a bug (or a feature?) in the syslog-ng daemon. I used a filter: filter f_host11 { host("192.168.1.11"); }; filter f_host111 { host("192.168.1.111"); }; destination d_host11 {file("/var/log/other/11.log");}; destination d_host111 {file("/var/log/other/111.log");}; log{source(s_other);filter(f_host11);destination(d_host11);}; log{source(s_other);filter(f_host111);destination(d_host111);}; Then, when I try to log something on 192.168.1.111, I receive a log record in both files: 111.log and 11.log too ! What am I doing wrong ? bye, Lubos. __________________________________________________ Do You Yahoo!? Sign up for SBC Yahoo! Dial - First Month Free http://sbc.yahoo.com

3 2

Solaris 2.6 + syslog-ng 1.5.18
by Sangbutsarakum, Patai 28 Jun '02

28 Jun '02

Dear lists, =1= I've tried to install 1.5.18 but without luck. My environment is Solaris 2.6 SPARC, I've success in 1.4.15 but I need TEMPLATE feature so I moved. To log to MySQL. Please suggest. =2= In case of tremendous of messages to log to MySQL via pipe method. Is that pipe tolerate for big load? Thanks in advances Pat. == Messages at Compile time == # ./configure creating cache ./config.cache checking for a BSD compatible install... ./install-sh -c checking whether build environment is sane... yes checking whether make sets ${MAKE}... yes checking for working aclocal... missing checking for working autoconf... missing checking for working automake... missing checking for working autoheader... missing checking for working makeinfo... missing checking whether build environment is sane... yes checking for gcc... gcc checking whether the C compiler (gcc ) works... yes checking whether the C compiler (gcc ) is a cross-compiler... no checking whether we are using GNU C... yes checking whether gcc accepts -g... yes checking for gcc option to accept ANSI C... none needed checking for bison... bison -y checking how to run the C preprocessor... gcc -E checking for flex... flex checking for flex... (cached) flex checking for yywrap in -lfl... yes checking lex output file root... lex.yy checking whether yytext is a pointer... yes checking whether make sets ${MAKE}... (cached) yes checking for ANSI C header files... yes checking for malloc.h... yes checking for unistd.h... yes checking for door.h... yes checking for stropts.h... yes checking for sys/strlog.h... yes checking for stdarg.h... yes checking for sys/klog.h... no checking for working const... yes checking whether time.h and sys/time.h may both be included... yes checking for modern utmp... yes checking for hosts_access in -lwrap... no checking whether to compile klogctl... no checking size of short... 2 checking size of int... 4 checking size of long... 4 checking for I_CONSLOG... yes checking for O_LARGEFILE... yes checking for working alloca.h... yes checking for alloca... yes checking for vprintf... yes checking for door_create in -ldoor... yes checking for select... yes checking for snprintf... yes checking for vsnprintf... yes checking for strerror... yes checking for inet_aton... no checking for strncpy... yes checking for getutent... yes checking for getopt_long... no checking for strcasecmp... yes checking for strptime... yes checking for pthread.h... yes checking for pthread_create in -lpthread... yes checking whether to enable Sun STREAMS support... yes checking whether to enable Sun door support... yes checking whether to enable TCP wrapper support... no checking libol version >= 0.3.3... ok updating cache ./config.cache creating ./config.status creating Makefile creating src/Makefile creating src/tests/Makefile creating doc/Makefile creating doc/sgml/Makefile creating contrib/Makefile creating utils/Makefile creating syslog-ng.spec creating src/config.h # make Making all in src cd /usr/local/src/syslog-ng-1.5.18 && /usr/local/src/syslog-ng-1.5.18/missing autoheader WARNING: `autoheader' is missing on your system. You should only need it if you modified `acconfig.h' or `configure.in'. You might want to install the `Autoconf' and `GNU m4' packages. Grab them from any GNU archive site. cd .. \ && CONFIG_FILES= CONFIG_HEADERS=src/config.h \ /bin/sh ./config.status creating src/config.h src/config.h is unchanged make all-recursive Making all in . gcc -DHAVE_CONFIG_H -I. -I/usr/local/src/syslog-ng-1.5.18/src -I. -g -O2 -Wall -I/usr/local/include/libol -D_GNU_SOURCE -c main.c main.c: In function `main': main.c:437: warning: passing arg 2 of `resolve_user_group' from incompatible pointer type main.c:437: warning: passing arg 3 of `resolve_user_group' from incompatible pointer type main.c:441: warning: passing arg 2 of `resolve_group' from incompatible pointer type gcc -DHAVE_CONFIG_H -I. -I/usr/local/src/syslog-ng-1.5.18/src -I. -g -O2 -Wall -I/usr/local/include/libol -D_GNU_SOURCE -c sources.c gcc -DHAVE_CONFIG_H -I. -I/usr/local/src/syslog-ng-1.5.18/src -I. -g -O2 -Wall -I/usr/local/include/libol -D_GNU_SOURCE -c center.c gcc -DHAVE_CONFIG_H -I. -I/usr/local/src/syslog-ng-1.5.18/src -I. -g -O2 -Wall -I/usr/local/include/libol -D_GNU_SOURCE -c filters.c filters.c: In function `make_filter_netmask': filters.c:314: warning: implicit declaration of function `inet_aton' gcc -DHAVE_CONFIG_H -I. -I/usr/local/src/syslog-ng-1.5.18/src -I. -g -O2 -Wall -I/usr/local/include/libol -D_GNU_SOURCE -c destinations.c gcc -DHAVE_CONFIG_H -I. -I/usr/local/src/syslog-ng-1.5.18/src -I. -g -O2 -Wall -I/usr/local/include/libol -D_GNU_SOURCE -c log.c gcc -DHAVE_CONFIG_H -I. -I/usr/local/src/syslog-ng-1.5.18/src -I. -g -O2 -Wall -I/usr/local/include/libol -D_GNU_SOURCE -c cfgfile.c gcc -DHAVE_CONFIG_H -I. -I/usr/local/src/syslog-ng-1.5.18/src -I. -g -O2 -Wall -I/usr/local/include/libol -D_GNU_SOURCE -c cfg-grammar.c cfg-grammar.y: In function `yyerror': cfg-grammar.y:703: warning: implicit declaration of function `fprintf' cfg-grammar.y:703: `stderr' undeclared (first use in this function) cfg-grammar.y:703: (Each undeclared identifier is reported only once cfg-grammar.y:703: for each function it appears in.) *** Error code 1 make: Fatal error: Command failed for target `cfg-grammar.o' Current working directory /usr/local/src/syslog-ng-1.5.18/src *** Error code 1 make: Fatal error: Command failed for target `all-recursive' Current working directory /usr/local/src/syslog-ng-1.5.18/src *** Error code 1 make: Fatal error: Command failed for target `all-recursive-am' Current working directory /usr/local/src/syslog-ng-1.5.18/src *** Error code 1 make: Fatal error: Command failed for target `all-recursive'

1 0

Syslog-NG security user permissions.
by Russo, Ben 27 Jun '02

27 Jun '02

With the recent brewhaha about SSH I can't help but wonder about other daemons running on my boxes that don't have privelage separation. Is it possible to give Syslog-ng a command line option (like named or ntpd)? So that after Syslog-ng binds to the network socket (with root privs) then it sets it's UID to something other than root? I realize that many of the options in syslog-ng might be more complex if this were done. I can think of many permissions and output file name and directory macros whose code would have to be modified if syslog-ng were to properly run as a regular user instead of ROOT and be able to properly handle error messages and such for permissions and directories and everything. However, it is inevitable with the facts that Syslog-NG is a network Daemon, that receives input and has macros based on that input to write to files, that a remote vulnerability in Syslog-NG will become known.... If Syslog-NG is running as a non-root UID then this is not a problem, (other than a big nuisance). -Ben.

2 1

Re: Thinking about linking
by Jay Goldberg 27 Jun '02

27 Jun '02

Hi Klaas Thanks very much for the reply and for thinking that this makes sense :) Symlinks are really portable - the symlink() and unlink() system calls are part of the Open Group's UNIX 98 definition (formerly X/Open and 'Spec 1170'), so any UNIX would be able to do this. Less sure about windows, BTW, although Cygwin and U/WIN definitely do support the calls. If I can find the time, I'll have a look at it myself, as I think it's really just a few lines of logic (plus error checking et al): add file options "symlink_cur" and "symlink_pre" at file close() time: if (symlink_cur is set) { unlink(symlink_cur) } if (symlink_pre is set) { symlink(real_file, symlink_pre) } at file open() time: if (symlink_cur is set) { symlink(real_file, symlink_cur) } ...that should about cover it. Never having touched lex & yacc, I've been hesitant to try to touch this code, but it shouldn't be that bad (famous last words :)) Kind regards Jay ===< Klaas van Gend wrote: >=== > From: klaas.vangend(a)philips.com > To: syslog-ng(a)lists.balabit.hu > Date: Wed, 26 Jun 2002 09:49:28 +0200 > Subject: [syslog-ng]Re: Thinking about linking > Reply-To: syslog-ng(a)lists.balabit.hu > > On 2002-06-26 06:25:05 AM syslog-ng-admin wrote: > > > > > Message: 4 > > Date: Tue, 25 Jun 2002 17:06:22 +0200 > > From: Jay Goldberg <jg(a)open.ch> > > To: syslog-ng(a)lists.balabit.hu > > Organization: Open Systems AG - http://www.open.ch > > Gecko/20020530 > > Subject: [syslog-ng]Thinking about linking > > Reply-To: syslog-ng(a)lists.balabit.hu > > > > Hi, syslog-ng'ers > > > > A quick question for all - I've been using syslog-ng, but instead of using > > the excellent built-in $YEAR $MONTH and $DAY macros for automatic logfile > > rotation, I've been using a different solution - > > > > I use the cronolog program (www.cronolog.org) as a destination for any files > > that I want to keep locally. It's not optimal, but it allows me to do one > > thing that keeps our complex management environment intact after the cutover > > from syslog to syslog-ng, and that's to maintain a symbolic link to the > > latest and previous logfiles in the place where folks/scripts would normally > > think to look for them. So for example: > > > > /reallybigdisk/logs/2002/06/messages.25 and > > /reallybigdisk/logs/2002/06/messages.24 > > > > would be today's and yesterdays 'messages' file. But with cronolog, I'm abl e > > to specify options so that /var/adm/messages is just a symlink to whatever f ile > > we're logging to now, and /var/adm/messages.0 is a symlink to whatever file > > we logged to last: > > > > host# ls -l /var/adm/messages* > > lrwxrwxrwx 1 root root 34 Jun 25 01:16 /var/adm/messages -> > > /reallybigdisk/logs/2002/06/messages.25 > > lrwxrwxrwx 1 root root 34 Jun 24 01:16 /var/adm/messages.0 -> > > /reallybigdisk/logs/2002/06/messages.24 > > > > I'm not sure if I missed this in the documentation for syslog-ng or perhaps > > I'm just insane for thinking that this is useful :) > > > > So I guess my questions after all this are: > > > > - did I miss this in the documentation? > > - am I insane for wanting such a feature? > > - would it be a lot of trouble to add this feature? > > > > ...I was thinking something along the lines of a 'maintain_symlink' option f or > > a file destination that takes the target of the symlink (e.g., > > /var/adm/messages) > > as an option. > > > As far as I know (and I've probably been the last person looking over that cod e), there is no such option currently available. > It wouldn't be too hard to write, if and only if creating symlinks is portable over all UNIX implementations. > Being a Linux&Windows programmer only, I don't know for sure. > Anyway, following the lines of my latest patch (the "remove_if_older" patch, s ubmitted Feb1 2002, included in 1.5.15), > it shouldn't be too difficult to implement. > > I currently don't have time for it, but I think your suggestion makes sense an d maybe I can persuade my boss to > give me some time to implement it. > > Groetjes, > > Klaas van Gend -- jay goldberg senior engineer open systems ag raeffelstrasse 29 ch-8045 zurich t: +41 1 455 74 00 f: +41 1 455 74 01 jg(a)open.ch http://www.open.ch

2 1