[Bug 430377] Review Request: atop - An advanced interactive monitor to view the load on system and process level

[bugzilla at redhat.com]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug report.

Summary: Review Request: atop - An advanced interactive monitor to view the load on system and process level


https://bugzilla.redhat.com/show_bug.cgi?id=430377





------- Additional Comments From wolfy at nobugconsulting.ro  2008-03-28 17:56 EST -------
First of all, I must take back the c) from comment #12. I was wrong, launching
scripts with "daemon" is perfectly valid. I am sorry for misleading you.

There are still a couple of small problems:
a) minimal importance: all the install lines include the option "-p" twice.
Please make sure you delete one of them before uploading the files to CVS
b) I suggest to replace the "start atop for all processes with interval of 10
minutes" with the more accurate "start atop for all processes with interval of
$INTERVAL (by default 10) minutes". I have no objections if you want to keep it
as it is, it's a just a cosmetic issue.
c) could you please explain the purpose of the section from cron.daily which
verifies if atop.log exists and creates it if it does not ? As far as I have
understood, the process goes like this (please correct me if I am wrong):
- the startup script launches atop as a daemon, writes raw information in a
/var/log/atop/atop_$currentdate file and redirects stdout and stderr to
/var/log/atop/atop.log (which gets either created or overwritten with this occasion)
- the daily cron job verifies if atop is running and 1) if it is running, forces
a flush of the current state, kills the daemon and restarts it with a new raw
log 2) exits without doing anything if atop is not running 3) deletes old logs.
  If I am correct
- I fail to see the need for checking at the start of the script if the logfile
exists. It gets created anyway when the daemon is restarted and forcing an
advanced re-creation of it is useless in the event that it was deleted while the
daemon was running.
- deleting old logs is usually a job for logrotate. Admins might wish to keep
logs for a longer time and deleting their logs is not a nice thing to do. On the
other hand, since each raw log file has a different filename, logrotate cannot
be used to rotate them. Therefore I will not object if you want to preserve the
"find ... -exec rm {}" line. However I suggest the following:
1) create two sections in the logrotate file you supply:
- existing one should be used for atop_* files only (which this way would be
archived daily)
- one with a longer lifetime ( rotate 4 + weekly) for atop.log
2) make a note somewhere (for instance a separate readme.fedora or a couple of
lines in sysconfig/atop ) that mentions that by default raw logs are deleted
after 28 days. You could even add another variable in sysconfig/atop and use it
in the find line in cron.daily, for instance:
        #the cron.daily job deletes raw data after 28 days
        DATALIFE=28
and
# delete logfiles older than four weeks
#
(sleep 3; find $LOGPATH -name 'atop_*' -mtime +$DATALIFE -exec rm {} \; )&



-- 
Configure bugmail: https://bugzilla.redhat.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.