/etc/weekly is run?
Rebuilding locate database: stty: stdin isn't a terminal TERM: Undefined
variable.
'su' apparently runs in root's environment before doing a seteuid(), so
it reads root's .cshrc. The problem lies in the fact that $prompt is
*always* set by the time the command is executed, so you can't protect
your interactive commands from the non-interactive commands by the
normal:
if ($?prompt) then
# interactive stuff here
endif
Set the -f flag (su -fm) which causes it not to read root's .cshrc file,
since there should be no reason to do so for these scripts since
/etc/crontab sets up all of the necessary environment.
"hand", changed /etc/crontab to call /usr/sbin/newsyslog every hour
(the entry was there before - but we haven't had any newsyslog until
today :-) and changed /etc/inetd.conf to also contain (commentet out)
entries for rpc.rquotad and rpc.sprayd (taken from NetBSD)
Add the `lc' option to default. Having getty understand the issues of
CAPS-only terminals is a nop these days, but `lc' avoids hassles for
people with CAPS-only user names.
Suggested by: Don Yuniskis
weekly:
Modify the call to makewhatis, now that our makewhatis can take the
entire ${MANPATH} as a single argument.
Leave a warning to the sysop if (s)he didn't yet
enable the /tmp cleanup code.
Made `core' in the cleanup template look `*.core'.
Replace `df -k' by `df -k -t local', since the stats
for kernfs, procfs etc. are not of much interest, and
the inclusion of nfs systems might hang the machine (nor is it a
`disk' statistic as the headline's telling).
weekly:
Modified the locate.updatedb part to work even if there's no
database yet; report errors other than `Permission denied' instead
of silently ignoring all of them.
Added functionality to rebuild the whatis database once a week.
