With openssh-server-2.9p2-3, this is what I see in /var/log/boot.log after sshd starts on reboot: Jul 22 15:42:55 jik sshd: Starting sshd: Jul 22 15:42:57 jik sshd: succeeded Jul 22 15:42:57 jik sshd: ^[[60G[ Jul 22 15:42:57 jik sshd: Jul 22 15:42:57 jik rc: Starting sshd: succeeded This is obviously not formatted correctly. I believe that the correct fix for this is to simply use "daemon" instead of using "initlog -c" directly from the sshd start-up script. I'll attach a patch to /etc/rc.d/init.d/sshd.
Created attachment 24547 [details] Use "daemon" instead of "initlog" when starting sshd
The daemon() function (in at least certain versions of initscripts, but not the current version) refuses to start a daemon if it detects it already running, so if you happen to be logged in via ssh when you try to start it up, it won't happen unless you call initlog directly. I'll make a note of this in the changelog so that I don't forget why it was changed to use initlog in the future. Marking as won't fix for this reason (you are correct that daemon() is generally the correct way to start daemons).
OK, I am confused. You seem to be saying that you can use daemon in the start-up script in the current version of sshd but not in previous versions. In that case, why don't you just fix the bug in Raw Hide rather than marking it WONTFIX? Or perhaps mark it DEFERRED instead of WONTFIX? In other words, I don't understand why you would mark a but WONTFIX when you've acknowledged that it should be fixed and *can* be fixed at some later date. Furthermore, even when using versions of initscripts where daemon won't work in this context, you can still fix this bug, i.e., get rid of the extra cruft in the boot log. I will attach another patch which shows how to do this.
Created attachment 47900 [details] Fix sshd start-up script boot log cruft without using daemon()
Different fix in initscripts-7.08-1.