From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jes Sorensen Subject: Re: [PATCH 1/3] Add support for launching mdmon via systemctl instead of fork/exec Date: Fri, 25 Jan 2013 16:05:31 +0100 Message-ID: References: <1358774578-2183-1-git-send-email-Jes.Sorensen@redhat.com> <1358774578-2183-2-git-send-email-Jes.Sorensen@redhat.com> <20130122164623.1962171b@notabene.brown> Mime-Version: 1.0 Content-Type: text/plain Return-path: In-Reply-To: (Jes Sorensen's message of "Fri, 25 Jan 2013 12:13:21 +0100") Sender: linux-raid-owner@vger.kernel.org To: NeilBrown Cc: linux-raid@vger.kernel.org, dledford@redhat.com, harald@redhat.com List-Id: linux-raid.ids Jes Sorensen writes: > Jes Sorensen writes: >> NeilBrown writes: >>> - I have come to the conclusion that --offroot is a bad idea. We should >>> just make that the default. No matter whether the array is providing the >>> root filesystem or not, I never want systemd (or anything else) to kill >>> mdmon. I want it to remain in control. So the systemctl handling should >>> assume offroot. e.g. there should only be one .service file. >> >> Ok I guess the question is what happens if an array is shut down in >> userland, does it take mdmon down manually once it is finished with it? >> We need this to happen, because otherwise we end up with a dangling >> mdmon process once we reboot, if the IMSM array wasn't assembled in the >> initrd. > > I have been digging some more into this. Basically it looks like we can > nuke the --offroot argument and simply default to running that way. > > On the bad side, we need two different .service files, one for on the > initrd and one for post system boot, due to different condition > requiremnts. The question is whether to name them the same, and just > have dracut know about the magic initrd version or if we want to try for > the two versions. Very odd - more testing and it looks like this may not be a problem after all, but instead because something went wrong during the creation of the initrd in my testing. It does make me a little concerned if this can happen to users who are installing an update, but with nothing more than a single incident there isn't much I can do. I will look at a set of patches. Jes