I believe the file /etc/init.d/boot.local is the SuSE equivalent of rc.local, at least boot.local is "where you can enter additional commands to be executed at boot time before changing to a run level" (cf 7.2 manual, page 313)
doug _______________________________________________
Thanks, Doug. I found that file, and it says the exact same thing at the end of the file. Any idea why the same file also exists in /etc/rc.d/ along with boot.setup and some others? Greg
On Thu, Feb 14, 2002 at 05:41:05PM -0500, doug waud wrote:
Gregory Avedissian wrote re SuSE's boot.local
Any idea why the same file also exists in /etc/rc.d/ along with boot.setup and some others?
Direct answer, no :-(
I did a locate boot.local and got back only /etc/init.d/boot.local
(I tried this on five machines here running SusE 7.2 and they all showed just one copy)
However, just to be sure, I went to /etc/rc.d and a ls -al gave me that boot.local you describe!
So now, I am at a loss as to why locate could not find it :-( but, anyway ...
The locate command looks at the database that updatedb generates nightly out of a cron job. The user that updatedb runs as probably doesn't have permission to list the contents of the /etc/init.d/ directory. -- Frank Sweetser fs at wpi.edu, fs at suave.net | $ x 16 Full-time WPI Network Tech, Part time Linux/Perl guy | Just don't make the '9' format pack/unpack numbers... :-) -- Larry Wall in <199710091434.HAA00838@wall.org>
Gregory Avedissian wrote re SuSE's boot.local
Any idea why the same file also exists in /etc/rc.d/ along with boot.setup and some others?
Direct answer, no :-( I did a locate boot.local and got back only /etc/init.d/boot.local (I tried this on five machines here running SusE 7.2 and they all showed just one copy) However, just to be sure, I went to /etc/rc.d and a ls -al gave me that boot.local you describe! So now, I am at a loss as to why locate could not find it :-( but, anyway ... I then poked around in the boot scripts (cf "SuSE Boot Concept" in the manual p 309 ff, or just man init.d ) and took a look at /etc/rc.d/boot --- the file init runs at the beginning. In there you can search for "local" (i.e. less /etc/rc.d/boot and then enter /local <return> followed by three N's each to move on to the Next place there is a "local" ) and see that init, the program that counts here, looks for and at /etc/init.d/boot.local. So this indicates it is the boot.local in /etc/rc.d which is the odd one. There also does not seem to be a symbolic link between them. With SuSE I am always suspicious of what SuSEconfig and /etc/rc.config are doing so I wondered whether SuSEconfig was looking at both boot.local files. But it seems to look at neither; in particular, it does not use one to update the other. I poked around on the web and, at http://www.linuxdoc.org/HOWTO/mini/Wacom-USB-mini-HOWTO-4.html found "Next I added in the /etc/rc.d/boot.local file (remember, I use a SuSE distribution, but you should fine an equivalent for RedHat in /etc/rc.d/init.d/boot.local or /etc/rc.d/init.d/rc.local)" This makes me think the /etc/rc.d location is atavistic, harking back to earlier versions of SuSE. I looked in my old 5.2 manual and could not see a clear proof of this but I do note that with the 7.2 SuSE version they seem to have tried to stick more closely to standards (page 287) so that old /etc/rc.d resting place may be in for some backwards compatibility with user's old scripts. My only problem with that theory is that I cannot find where the /etc/rc.d/boot.local file would get read in 7.2 The bottom line is clear --- use /etc/init.d/boot.local if you want to add a special modprobe or the like in recent SuSE distros. But, I'll still keep my eyes open for confirmation of the above or an alternative view. doug
Frank Sweetser wrote:
So now, I am at a loss as to why locate could not find /etc/rc.d/boot.local
The locate command looks at the database that updatedb generates nightly out of a cron job. The user that updatedb runs as probably doesn't have permission to list the contents of the /etc/init.d/ directory.
Made sense to me then, like a fool, I tried running updatedb as root and then, in belt-and-suspenders mode, I tried updatedb --localuser=root but still got only the boot.local in /etc/init.d :-( I guess this is why you let sleeping dougs lie doug
I did a locate boot.local and got back only /etc/init.d/boot.local
(I tried this on five machines here running SusE 7.2 and they all showed just one copy)
However, just to be sure, I went to /etc/rc.d and a ls -al gave me that boot.local you describe!
Hi Doug, The reason that /etc/init.d/boot.local and /etc/rc.d/boot.local are the same file is because /etc/rc.d is a link to /etc/init.d on my SuSE 7.2 PRO system. linuxpc> ls -ald /etc/rc.d lrwxrwxrwx 1 root root 6 Mar 24 2001 /etc/rc.d -> init.d Later, Andy -- Andy Stewart Founder Worcester Linux Users' Group Worcester, MA, USA http://www.wlug.org
On Sat, Feb 16, 2002 at 12:02:40PM -0500, doug waud wrote:
I then went back to an old UNIX manual and did some more reading :-) and tried
root@233p5:/ > ls -il /etc/init.d/boot.local 274722 -rwxr--r-- 1 root root 447 Jan 30 16:37 /etc/init.d/boot.local root@233p5:/ > ls -il /etc/rc.d/boot.local 274722 -rwxr--r-- 1 root root 447 Jan 30 16:37 /etc/rc.d/boot.local
and, miracles still happen, got the same inodes for both files. So it looks like we have a hard link. However, we still have 1 and not 2 :-)
Aha - there's the confusion. There's no problem here - boot.local only has one directory entry (aka hardlink) pointing to it. /etc/init.d/ is a directory which contains a directory entry that maps the name boot.local to inum 274722. /etc/rc.d/ is *not* a directory, but rather a symlink. symlinks do not contain any information about that to which they point except the name, so it has no idea if it's pointing to a file, directory, dev entry, etc. So, when /etc/rc.d/foobar is opened, the symlink is hit and replaced with the contents of it's target -> /etc/init.d/foobar, and the entry foobar is looked up in the directory /etc/init.d/ Make sense? -- Frank Sweetser fs at wpi.edu, fs at suave.net | $ x 16 Full-time WPI Network Tech, Part time Linux/Perl guy | Tactical? TACTICAL!?!? Hey, buddy, we went from kilotons to megatons several minutes ago. We don't need no stinkin' tactical nukes. (By the way, do you have change for 10 million people?) --lwall
The reason that /etc/init.d/boot.local and /etc/rc.d/boot.local are the same file is because /etc/rc.d is a link to /etc/init.d on my SuSE 7.2 PRO system.
linuxpc> ls -ald /etc/rc.d lrwxrwxrwx 1 root root 6 Mar 24 2001 /etc/rc.d -> init.d
Been there, done that :-) Actually, I tried root@233p5:/ > ls -al /etc/init.d/boot.local -rwxr--r-- 1 root root 447 Jan 30 16:37 /etc/init.d/boot.local root@233p5:/ > ls -al /etc/rc.d/boot.local -rwxr--r-- 1 root root 447 Jan 30 16:37 /etc/rc.d/boot.local | |____________________ | and found that both had the number 1 where I expected 2. I then went back to an old UNIX manual and did some more reading :-) and tried root@233p5:/ > ls -il /etc/init.d/boot.local 274722 -rwxr--r-- 1 root root 447 Jan 30 16:37 /etc/init.d/boot.local root@233p5:/ > ls -il /etc/rc.d/boot.local 274722 -rwxr--r-- 1 root root 447 Jan 30 16:37 /etc/rc.d/boot.local and, miracles still happen, got the same inodes for both files. So it looks like we have a hard link. However, we still have 1 and not 2 :-) I notice you have the same "disease" I then wondered whether SuSE was doing something odd. I therefore went to /tmp and, as root, created two subdirectories z1 and z2. I then cat'd /etc/fstab to z1/test1 to create a file. I finally created a hard link from z2/test1 to z1/test1. When I now did a ls -il on those two test1 files, there were 2's where they should be! At this point, Chuck suggested I do a forced fsck. I had some trouble at first (until he showed me I had to remount / ro; I had tried the SuSE rescue system but it complained about bad superblocks and/or odd "magic numbers". This happened on three boxes so I was unconvinced there was a screwed up file system on all three machines). I have just now finished that fsck -f /dev/hda7 on my notebook and all it reports is 0.2% non-contiguous --- as I suspected from the problem on three boxes (or 4, if we include yours) no screwed up file system. So the mystery now is why do we both get 1 for those linked files. doug filesystem.
doug waud wrote: too much :-) I thought I was writing to Andy and somehow wasted everyone's time Meanwhile, Frank Sweetser has summarized the situation well. doug
participants (4)
-
Andy Stewart
-
doug waud
-
Frank Sweetser
-
Gregory Avedissian