I initially asked 3 questions in one email to the list.
Due to the large amount of information / responses, I have split
the summaries up into 3 separate entities.
Huge thanks and respect go out to all of the OSF-managers who
generously took the time to help. Thanks go to Erik Persson,
Larry Church, Subir Grewal, Barb Baker, Mikel Stous, Karl Marble,
alan_at_nabeth.cxo.dec.com, paw_at_northstar.dartmouth.edu and
Anil Khullar . I do hope I have not missed anyone out.
As others have mentioned, it is truly amazing how fast
replies / answers come in - OSF-managers is an awesome
list and this system of summaries (to keep traffic low)
is ideal.
Thanks again. (the original question and summary follow)
Chas
The original question was :
> 2) I am trying to track down all of the files that will
> grow and have the following :
> /var/adm/*
> webserver log files
> mail spool directories
> Have I missed anything else ?
Summary :
---------
1. Generalize that to include the whole /var - thats the place where
well-behaved programs put their logs and stuff.
[Erik Persson pretty much summed it up there.]
Everything in /var is liable to grow as the var is
short for variable. The purpose was to isolate all
the read/write files from the read-mostly files that
are the rest of /usr. Well behaved layered products
will put the right stuff in /var, but some programs
may spread things around where they feel like it.
[alan_at_nabeth.cxo.dec.com qualified it]
2. The location of the webserver and mail spools depend on your
own configuration.
3. [ Subir Grewal had some good advice too :]
That's about it. Nothing else really has the potential of getting huge,
unless your users are doing wierd things with crontab. Set up scripts
to
clear out /tmp every so often, and most sysadmins managing systems with
very large user bases set up scripts to look for large files with names
like core that haven't been accessed in a while. It's also wise to look
for the files that editors leave behind when they die unexpectedly, like
#pine*
4. [ Karl Marble found one more ]
/usr/lib/cron/log
5. [ Larry Church pointed me to yet another good resource ]
There are (or can be) a gizillion log files. Most Digital Unix (DU) log
files are set up to use the syslog.dated facility. See the syslogd man
page for details.
NeWT (Neuronet Web Team)
http://heaven.com.my - Channel X.
http://neuronet.com.my - Home.
Received on Sat Mar 29 1997 - 13:27:42 NZST