It turns out that the stock kernel 2.4.20 from www.kernel.org doesn't have this problem. -----Original Message----- From: Michael Daskalov Sent: Thursday, May 08, 2003 10:29 PM To: suse-linux-e@suse.com Subject: Problems wih software raid. Hi all, I have SuSE 7.2 on a machine with 4 HDDs. I have configured my root and boot to be on RAID1. I installed the machine with he latest available kernel as a patch to SuSE 7.2 (ftp://ftp.suse.com/pub/suse/i386/update/7.2/kernel/2.4.18-20030324/k_deflt-2.4.18-262.i386.rpm) I had a problem with the IDE devices. Two blocks on two different disk were reported as bad and my raid5 was broken. Testing the IBM 120GB disks with the software provided by IBM/Hitachi revealed there were no problems in the disks. Hoping that this was due to a problem with the IDE driver in the last kernel, I tried ftp://ftp.suse.com/pub/people/mantel/next/RPM/kernel-source-2.4.20.SuSE-62.i586.rpm I compiled it myself including md,raid1,raid5 in the kernel (not as modules) and there exists the following problem: After every shutdown or reboot my /dev/md0 (rootfs) device needs recyncing. It's because it's not stopped cleanly on shutdown or reboot. All other /dev/md? devices are stopped on shutdown, only this one is not. On the console I see messages like ....skip.... /dev/md1: switching to read-only mode ....skip.... /dev/md2: switching to read-only mode ....skip.... /dev/md3: switching to read-only mode ....skip.... /dev/md4: switching to read-only mode but I don't see that message for /dev/md0. Instead I see a message saying that /dev/md0 is still in use. Anyone know why this happens on newer kernels? How can I fix the problem? I didn't have this problem on k_deflt-2.4.18-262.i386.rpm . Does this version have any known problems with Intel DG845 motherboards? Help me please, Mihail Daskalov MDaskalov@technologica.biz