Red Hat Bugzilla – Bug 58767
RH 6.2 updated atd initscript fails
Last modified: 2007-04-18 12:39:10 EDT
From Bugzilla Helper:
User-Agent: Mozilla/4.77 [en] (X11; U; Linux 2.2.19-kim.1 i486)
Description of problem:
RHSA-2002:015-13 recommends at-3.1.8-22.1.i386.rpm as an updated at package for
use with Red Hat 6.2 systems.
However, the atd initscript installed into /etc/rc.d/init.d/atd by the above
update is not consistent with the initscripts-5.00-1.i386.rpm package used in RH
IOW, after the update, 'service atd restart' fails.
The /etc/rc.d/init.d/atd script contains:
# Source function library.
In RH 6.2 systems that file should be: /etc/rc.d/functions
In addition, the updated atd script contains lines such as:
echo -n $"Starting $prog: "
RH 6.2 systems do not support internationalized startup messages, so the above
shows up with a spurious "$".
Version-Release number of selected component (if applicable):
Steps to Reproduce:
Use the source.
Workaround is to the atd initscript from at-3.1.7-14.i386.rpm.
However, this workaround is not consistent with the following from 3.1.8-22.1:
postuninstall scriptlet (through /bin/sh):
if [ "$1" -ge "1" ]; then
service atd condrestart >/dev/null 2>&1
because the 3.1.7-14 atd script does not support 'condrestart'.
The stop() function in the same file also contains the line:
I cannot find any package providing this in redhat 6.2.
Doing 'ln -s rc.d/init.d /etc/init.d' appears at first glance the easiest way to
make atd start.
Rather than adding a symlink which will not be upgraded via the initscripts
package, just edit /etc/rc.d/init.d/atd and correct the location of functions.
This package upgrade was not properly tested before being released since this is
a fairly blatent error. I hope we don't see more of these problems in the future :-(
Red Hat 7.1 and 7.2 systems contain the /etc/init.d link ("cd /etc; ln -s
rc.d/init.d init.d"), thus adding the link is not an inconsistent way to handle
On the 7.x systems the link is provided by the chkconfig RPM. (The entire
/etc/rc?.d directory structure is laid out via the chkconfig RPM *and* the
initscripts RPM, but the /etc/init.d soft link exists only in chkconfig.
Hmmm???) In any case, Red Hat could fix the problem by leaving the "at" RPM in
place and issuing a new initscripts or chkconfig RPM for 6.2 that contains the
/etc/init.d link. Of course, this does not address the '$' on the echo, but at
least the script will work and the system will be ready for other init.d scripts
that expect the /etc/init.d link!
This is also broken because of requiring sendmail, maybe the package should check for other MTA's or simply make sure that the command sendmail exists so MTA's like postfix will pass the requires stage.
Created attachment 44821 [details]
patch to fix 3.1.8-22.1 for RH62 operation
I also have this problem on my Redhat 6.2 after upgarding to the new at rpm . I
don't know enough to know what the implications of the error are. I'm going to
downgrade until a new rpm is released.
By the way, a spurious $ also occurs when the Redhat Network (daemon?) is
( A Desktop Redhat Linux User since RH 5.0 )
at-3.1.8-22.2.i386.rpm has an updated atd initscript which correctly sources
Reference: Advisory ID: RHSA-2002:015-15 Updated on: 2002-02-07
Newly updated atd initscript still echos spurious '$', but I suppose this allows
a clean upgrade path
to internationalization support. And minor "prettiness" issue doesn't affect
AFAICT, this bug is ready to be closed (ERRATA).
Closed. See previous comment.