The usual system to use for log file rotation on Linux is logrotate the man page even includes an example entry for dealing with apache log file rotation.
Note that the logrotate example does a:
of the apache PID, not a restart of apache, perhaps you could try that in your savelog configuration.
The other idea to try is to do a copy truncate, that is copy the current log file to httpd.log.0 (having rolled the previous log files to httpd.log.n+1 first) and then do :
So that the inode number used by apache for the log file does not need to change?
Last edited by TonyFullerMalv; 04-02-2010 at 09:26 AM..
Reason: Extra ideas
Hello all.
Due to some reason I can not use HUP to rotate needed log files.
So I use the standard method:
cp $file $file.1
cat /dev/null > $file
But if Java application in this time writing the output to $file,
in the beginning of it appears many "^@^@^@^@^@^@".
How to avoid it? Or how... (6 Replies)
I have the below script to help with disk space cleanup that finds logs older than a specified number of days (say 10 days). I need it to grab "active" logs as well. Problem is an "active log" will not get archived unless I put in 0 days which I don't want to do, I need to leave the past 10 days,... (2 Replies)
Hi, I current have many apps servers running and need to create a script to rotate logs daily, and then create an autosys job to delete logs that are older than 30 days. I was thrown into this and have no idea what to do, please help me get started, thanks!
-----Post Update-----
and i will... (6 Replies)
Hi , I have cronolog (http://cronolog.org/) setup to rotate both my error and access logs, but how can I use it to rotate my mod_jk.log file. I tried different things but cant get it to work. I want to end up where all three files will be rotated daily in YYYY/MM/DD folder format. thanks (0 Replies)
Hi
I am trying to create a simple function that implements log rotation
rotatelogs () {
file_name=$1
archive_dir=$2
retention_period=$3
find $archive_dir -name $file_name -mtime +$retention_period -exec rm -f {} \;
}
Issue i am facing is the file_name would be something like my... (3 Replies)
hi folk,
need advise regarding the log rotation, i have the logadm set at
30 2 * * * /usr/sbin/logadm
so it supposed to rotate once per day, but now it rotated twice!
but someone my log will rotate at 2:30 AM, but then another 2 hours later, it creates a new and rotate a new log again,... (2 Replies)
Hi Guys,
Good morning, I just want to know and collect ideas on this one. Regarding rotation of logs as I've observed it's not consistently functioning. I have a server with 8 Partitions, each partition has a dedicated directory for the logs that is needed and I set it every 5mins (300secs) the... (1 Reply)
Need some suggestion with an old problem form a thread here:
https://www.unix.com/shell-programming-scripting/153204-multiple-not-statement-awk.html
Since my log is large, I did get help to make a line that splits the log into two partstail -f syslog | awk '!/snmpd|ntpd|reject/{print | "tee... (1 Reply)
Hi All!
I seem to have a problem with log rotation, unless I am doing something wrong, I have type the following command for testing purposes to see if the -s option works but he did not:
logadm -w /var/adm/messages -C 8 -c -s 512k -t '/var/adm/messages.$n' -z 1
the file is now at this... (7 Replies)
Discussion started by: fretagi
7 Replies
LEARN ABOUT DEBIAN
mergelogs
MERGELOGS(1) General Commands Manual MERGELOGS(1)NAME
mergelogs - merge and consolidate web server logs
SYNOPSIS
mergelogs -p penlog [-c] [-d] [-j jitter] [-t seconds] server1:logfile1 [server2:logfile2 ...]
EXAMPLES
mergelogs -p pen.log 10.0.0.1:access_log.1 10.0.0.2:access_log.2
mergelogs -p pen.log 10.0.18.6:access_log-10.0.18.6 10.0.18.8:access_log-10.0.18.8
DESCRIPTION
When pen is used to load balance web servers, the web server log file lists all accesses as coming from the host running pen. This makes it
more difficult to analyze the log file.
To solve this, pen creates its own log file, which contains the real client address, the time of the access, the target server address and
the first few bytes of the requests.
Mergelogs reads pen's log file and the log files of all load balanced web servers, compares each entry and creates a combined log file that
looks as if the web server cluster were a single physical server. Client addresses are replaced with the real client addresses.
In the event that no matching client address can be found in the pen log, the server address is used instead. This should never happen, and
is meant as a debugging tool. A large number of these indicates that the server system date needs to be set, or that the jitter value is
too small.
You probably don't want to use this program. Penlog is a much more elegant and functional solution.
OPTIONS -c Do not cache pen log entries. The use of this option is not recommended, as it will make mergelogs search the entire pen log for
every line in the web server logs.
-d Debugging (repeat for more).
-p penlog
Log file from pen.
-j jitter
Jitter in seconds (default 600). This is the maximum variation in time stamps in the pen and web server log files. A smaller value
will result in a smaller pen log cache and faster processing, at the risk of missed entries.
-t seconds
The difference in seconds between the time on the pen server and UTC. For example, this is 7200 (two hours) in Finland.
server:logfile
Web server address and name of log file.
AUTHOR
Copyright (C) 2001-2003 Ulric Eriksson, <ulric@siag.nu>.
SEE ALSO pen(1), webresolve(1), penlog(1), penlogd(1)
LOCAL MERGELOGS(1)