|
From: Kris D. <kd...@vi...> - 2006-06-06 19:32:25
|
and...@fr... wrote: > No, the directory is correct : /etc/init.d/rc3.d *blink* *roll eyes* Just what we need; another SysV variant. :/ [snip] > if I look at the directory /etc/init.d/rc3.d i see > > sogeaste01:/etc/init.d/rc3.d # dir | grep amp > lrwxrwxrwx 1 root root 11 Jun 6 18:06 K15amportal -> ../amportal > lrwxrwxrwx 1 root root 11 Jun 6 18:06 S07amportal -> ../amportal *blink* Bizarre. I've *never* seen a system with both "start" (Snn) and "kill/stop" (Knn) links in a single rc?.d directory. > sogeaste01:/etc/init.d/rc3.d # dir | grep my > lrwxrwxrwx 1 root root 8 Jun 6 18:06 S12mysql -> ../mysql > > so now the numbers are "correct", no more 99; and also the link bacame > relative; If I reboot the server the job amportal at last starts. Completely Wild-Assed Guess: SuSE has added another level of indirection somewhere (init scripts actually in /etc/init.d/init.d maybe?), and its init system ignores any init script not actually in that directory. That doesn't entirely explain the failure of Webmin to create "proper" entries; I've got a number of systems with absolute syslinks in rc?.d directories and the service they refer to works fine. > So my problem is solved, thanks to you who suggested to use a "System > dependant" tool to adjust runlevel; > > but If now I use webmin to create a new bootup and shutdown action, let's > say aaa, again: > > sogeaste01:/etc/init.d/rc3.d # dir /etc/init.d/rc3.d | grep aaa > lrwxrwxrwx 1 root root 15 Jun 6 18:13 S99aaa -> /etc/init.d/aaa > > absolute link and 99 again... What about the other rc?.d directories? I tried a test on my server (White Box 3), and Webmin created a "proper" initscript (including the bare minimum entries for chkconfig - the shell-level init system tool used by RedHat and most of its close derivatives) and all of the proper symlinks (probably by using chkconfig). I didn't reboot, but I've *NEVER* seen a system service fail to start (or even throw up an error) if the symlinks were in place. -kgd |