Lighttpd rotating log files with logrotate tool

by on July 8, 2006 · 15 comments· LAST UPDATED December 18, 2007

in , ,

Lighttpd logo

Last time I wrote about setting up virtual hosting for Lighttpd web server. Naturally next step is to setup log rotating with logrotate which rotates, compresses log files.

Our setup

Our sample setup has total 6 log files:
Default domain/IP log files:
/var/log/lighttpd/access.log
/var/log/lighttpd/error.log

nixcraft.com virtual domain log files:
/var/log/lighttpd/nixcraft.com/access.log
/var/log/lighttpd/error.log

theos.in virtual domain log files:
/var/log/lighttpd/theos.in/access.log
/var/log/lighttpd/theos.in/error.log

logrotate Configuration

All you need to do is open/create logrotate configuration file for lighttpd. Open file /etc/logrotate.d/lighttpd:
# vi /etc/logrotate.d/lighttpd

Append following text:
"/var/log/lighttpd/*.log" "/var/log/lighttpd/nixcraft.com/*.log " "/var/log/lighttpd/theos.in/*.log " {
missingok
copytruncate
rotate 7
compress
notifempty
sharedscripts
postrotate
/etc/init.d/lighttpd reload
endscript
}

Where,

  • "/var/log/lighttpd/*.log" "/var/log/lighttpd/nixcraft.com/*.log " "/var/log/lighttpd/theos.in/*.log ": Log files with wild card specification as per our setup.
  • missingok: If the log file is missing, go on to the next log file without issuing an error message.
  • copytruncate: Truncate the original log file to zero size in place after creating a copy, instead of moving the old log file and optionally creating a new one
  • rotate 7: Log files are rotated 7 times before being removed or mailed to the address specified in a mail directive. If count is 0, old versions are removed rather then rotated.
  • compress: Old versions of log files are compressed with gzip to save disk space.
  • notifempty: Do not rotate the log if it is empty
  • sharedscripts
    postrotate
    /etc/init.d/lighttpd reload
    endscript:
    The lines between postrotate and endscript (both of which must appear on lines by themselves) are executed after the log file is rotated. These directives may only appear inside a log file definition. In our case we are reloading lighttpd. Other opting could be send –HUP single using kill command.

Make sure crond runs automatically after system reboot

Now your logs will rotate with logrotate command which is called from cronjob (/etc/cron.daily/logrotate) everyday. So make sure crond is running all the time:
# /etc/init.d/crond start
# chkconfig --list crond
# chkconfig crond on

Alternatively, run text based GUI tool for same purpose (Redhat/CentOS/Fedora and friends):
# ntsysv

If you are using Debian Linux, type the following command to configure crond using text based GUI tools:
# rcconf

Alternatively you can use update-rc.d command (Debian / Ubuntu Linux) to start crond automatically after system reboot:
# update-rc.d crond defaults

TwitterFacebookGoogle+PDF versionFound an error/typo on this page? Help us!

{ 15 comments… read them below or add one }

1 Colin January 11, 2007 at 12:28 am

If you are running BSD, you can use BSD’s newsyslog to do the same. Just add this line to /etc/newsyslog.conf:

/var/log/lighttpd.access.log \
644 7 * $D0 \
B /var/run/lighttpd.pid

It rotates the lighttpd access log everyday at midnight, saves seven previous logs, and sends -HUP to lighttpd. See “man newsyslog.conf” for more information.

Reply

2 Colin January 11, 2007 at 1:29 am

Oops, left out the owner:group parameter:

/var/log/lighttpd.access.log www

Reply

3 Colin January 11, 2007 at 1:31 am

Sigh… it got truncated:

/var/log/lighttpd.access.log www:www 644 7 * $D0 B /var/run/lighttpd.pid

(If you have lighttpd running as a different user/group than www, modify accordingly.)

Reply

4 nixCraft January 11, 2007 at 12:13 pm

Colin,

Thanks for sharing newsyslog configuration directive

Appreciate your post.

Reply

5 aman August 4, 2007 at 12:18 pm

What is this crond above refers to if u please tell me then its really helpful to me

Reply

6 F4jr September 2, 2008 at 9:48 am

Is it possible to do something like that :

“/var/log/lighttpd/*/*.log”

thanks for your help !

Reply

7 Rob October 28, 2008 at 1:30 am

@ F4jr
It sure is! You do need to tell newsyslog that the logfile is a shell pattern with the G flag:


/var/log/lighttpd.*.log www:www 644 7 * $D0 GJ /var/run/lighttpd.pid

Reply

8 F4jr china November 12, 2008 at 10:59 am

It works perfectly !

Thanks !

Reply

9 Geoff January 5, 2009 at 12:23 am

Why does lighttpd need to be restarted after rotating the log files? Isn’t the log simply copied, then truncated? I don’t see why a restart is needed.

Reply

10 Rytis September 15, 2009 at 2:32 am

Good question Geoff.

I also wonder why does lighttpd has to be restarted after logs are rotated?

If the logs are rotated everyday, the server is restarted every day.

Reply

11 Rob September 15, 2009 at 11:32 pm

@ Geoff and Rytis:

I don’t think the HUP signal actually restarts lighttpd, from my understanding it tells the daemon to close any open file handles, open a new logfile and reload the config file

Reply

12 Rytis September 16, 2009 at 12:49 am

@ Rob

mod_status reports lighttpd started up after logs were rotated, and all the previous stats are cleared. Since some of my logs are roated daily, my server uptime reported by mod_status is never more than a day.

I was just wondering if it’s possible to rotate the logs and keep the server running (or is it good to restart it every day anyways?)

Reply

13 Rob September 16, 2009 at 2:33 am

@ Rytis

mod_status shows no change in uptime between log rotations for me … lighttpd-1.4.23 on FreeBSD 7.2-STABLE

/etc/newsyslog.conf

/var/log/lighttpd.*.log www:www 600 7 * $D0 GJ /var/run/lighttpd.pid

Reply

14 Ela October 1, 2009 at 3:02 pm

Am getting this problem very biggly

Please help me out

vi /etc/logrotate.conf
# see “man logrotate” for details
# rotate log files weekly
weekly

# keep 4 weeks worth of backlogs
rotate 4

# create new (empty) log files after rotating old ones
create

# uncomment this if you want your log files compressed
#compress

# RPM packages drop log rotation information into this directory
include /etc/logrotate.d

# no packages own wtmp — we’ll rotate them here
/var/log/wtmp {
monthly
minsize 1M
create 0664 root utmp
rotate 1
}

cat /etc/logrotate.d/httpd

/var/log/httpd/*log {
missingok
notifempty
sharedscripts
postrotate
/sbin/service httpd reload > /dev/null 2>/dev/null || true
endscript
}

My server is running nearly 10 sites and all those creating log files under statistics in vhosts.
it wil become more than 30 gb with in week. how to control it. please reply me

Reply

15 Sagan September 21, 2012 at 12:03 pm

What if your Debian doesn’t have a crond?

Reply

Leave a Comment

Tagged as: , , , , , , , ,

Previous post:

Next post: