From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dave Young Subject: Re: multipath device issue Date: Fri, 03 Aug 2012 16:55:22 +0800 Message-ID: <501B91FA.6010703@redhat.com> References: <501A0B87.9070506@redhat.com> <501A3532.7010907@redhat.com> <501A7590.3010701@redhat.com> <501A7E2A.4020601@redhat.com> <501A8078.1020208@redhat.com> <501B635B.1090703@redhat.com> <501B7FCE.1050502@redhat.com> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <501B7FCE.1050502-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org> Sender: initramfs-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: text/plain; charset="us-ascii" To: Harald Hoyer Cc: "initramfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org" , Chao Wang On 08/03/2012 03:37 PM, Harald Hoyer wrote: > Am 03.08.2012 07:36, schrieb Dave Young: >> Hi, harald >> >> It has nothing to do with the udev rules. >> >> There's two issues: >> 1) 90multipath install multipath.sh to pre-trigger hook, but seems >> /dev/mapper/mpathX depends on /dev/dm-* population which will be created >> after udev trigger. >> >> 2) fstab-sys call wait_for_dev /dev/mapper/mpathb in initqueue, >> /dev/mapper/mpathb does not geting change to create, so dracut drop to >> emergency shell. >> >> I'm not sure what's the right way to resolve this, move multipath.sh to >> pre-mount hook looks resolve the issue 1), how to resolve the issue 2)? >> comment out it works, but it looks not a right fix. >> >> I'm not familiar with multipath, please correct me for anything wrong. >> > > I think it _has_ to do with the udev rules. What I should say is "It has nothing to do with the udev rules patches you mentioned previously" ;) > > 1. the multipathd daemon has to be started before udevd is started. > 2. when udev is triggered, 62-multipath.rules should kick in and call > "multipath" for every disk, which adds disks to the daemon, if they fit. > 3. of course /etc/multipath.conf has to be configured correctly. > > So, if you see /dev/mapper/mpathb be correctly created, when you start multpathd > after all devices exist in /dev, then the 62-multipath.rules > mechanism does not seem to work, > > With the next systemd release "rd.udev.log-priority=debug" does work again and > we might be able to debug this issue. Ok, will see if I can get something from the 62-multipath.rules. Thanks for the suggestion. -- Thanks Dave