Red Hat Bugzilla – Bug 51335
initscripts should depend on SysVinit
Last modified: 2014-03-16 22:22:24 EDT
Description of Problem:
initscripts' /etc/init.d/functions uses pidof, which is in SysVinit.
However initscripts does not depend on SysVinit either directly or through
I need to build rpms from time to time, and do this on a Debian system
inside a chroot. When I build an rpm that needed to start a service, and
tested it inside this chroot, startup failed because it didn't have SysVinit.
How to reproduce:
Here's the list of packages I have installed to build the package:
MAKEDEV autoconf automake basesystem bash binutils bzip2 bzip2-devel
chkconfig cpp cracklib cracklib-dicts db1 db2 db3 dev diffutils e2fsprogs
ed file filesystem fileutils findutils gawk gcc gcc-c++ gdbm glib glibc
glibc-common glibc-devel gnupg grep gzip info initscripts kernel-headers
libstdc++ libstdc++-devel libtermcap libtool libtool-libs logrotate m4 make
mingetty mktemp modutils mount ncurses net-tools pam passwd patch perl popt
procps psmisc pwdb rpm rpm-build rpm-devel sed setup sh-utils shadow-utils
slang strace sysklogd tar termcap textutils util-linux vim-common
vim-minimal which words zlib zlib-devel
None of these depend on SysVinit. I installed my package, it went ok,
started the service, worked fine, stopped the service, failed because it
couldn't find pidof.
I know this isn't the most common situation, but just think of someone who
does actually use Redhat, and wishes to use a chroot for some services as a
security measurement. He will run into the same problem.
PS. How acceptive are you to new packages? I generally don't care about
things I package for Redhat systems, as they are usually to fullfill very
specific needs, but this one is rather general-purpose and I think it could
be a useful addition.
Will be fixed in 6.13-1; thanks for the feedback!