Bug#310368: Bug #310368: MD not working...?

[Paul Check]

I am having similar problems as reported in Bug 310368, although I'm not certain if it is mdadm, my setup, or something else (like udev) that is causing my difficulties. General Setup: I am running Debian unstable, with the latest packages, on a 2.6.11 smp kernel, although I have a non-smp 2.6.10 kernel and an smp 2.4.26 kernel I use as backups. (So, if there is any possibility that constructing arrays in one system and deleting them in another, this could be a factor in my problem). I use initrd in both 2.6 kernels but not the 2.4 kernel. RAID Setup: I am running a raid 1 array, with devices md0, md1 and md2. My 2.6 kernels have raid 5 compiled in, but I don't have any raid 5s setup (this absolutely should not be a problem?). My setup is: md0: /dev/sda1 /dev/sdb1   -->  /boot md1: /dev/sda3 /dev/sdb3   -->  / md2: /dev/sda4 /dev/sdb4   -->  lvm, for other partitions. Note, I also have disks sdc through sdg, no raid, but using lvm. Problem: Yesterday after re-booting the system, the md1 and md2 devices were not getting created properly on boot. However, if I boot into the 2.4 kernel everything worked fine. After reading the bug report, I tried creating the /etc/mdadm/mdadm.conf file (I never had one before..my backup system shows that). With the mdadm.conf file created, I am now able to boot into my system in the 2.6 kernels: The md1 and md2 devices get created. However, for some reason, /dev/sda4 now won't get added to md2 (either automatically or manually). Again, if I boot into the 2.4 kernel, I can create everything, no problem. If I try to add /dev/sda4 to /dev/md2 I get: mdadm /dev/md2 -a /dev/sda4: mdadm: hot add failed for /dev/sda4: Invalid argument in dmesg: md: could not bd_claim sda4. md: error, md_import_device() returned -16 Note: I may have switched over to 2.4 and done some raid work after this problem arose in an attempt to fix it, but I presume that this shouldn't (?) damage the raid on the 2.6 kernels? On the other hand, I have had problems several months ago running lvm across the different kernels. This problem disappeared, but perhaps it's back and interferring with the sda4 partition somehow? (I doubt it...). Observation: The differences between my 2.6 kernels (raid not working) and 2.4 kernel (raid working) are: udev vs. non-udev and initrd vs none. I have had problems with udev in the past. I've also seen an old post regarding a bug with mdadm and the use of initrd's. I'm not sure I need my initrd's, so I'll try booting with that on the 2.6.10 non-smp kernel to see if the problem goes away. As for udev: ---1 dpkg -l udev: Desired=Unknown/Install/Remove/Purge/Hold | Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) ||/ Name                Version             Description +++-===================-===================-====================================================== ii  udev                0.056-2             /dev/ management daemon ---2 ls -ld /dev/.udevdb: drwxr-xr-x  2 root root 14560 2005-05-25 03:28 /dev/.udevdb Finally, here is a little other info: ---1 grep '/dev tmpfs' /proc/mounts: none /dev tmpfs rw 0 0 ---2 dpkg -l mdadm: Desired=Unknown/Install/Remove/Purge/Hold | Status=Not/Installed/Config-files/Unpacked/Failed-config/Half-installed |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) ||/ Name Version Description +++-===================-===================-====================================================== ii mdadm 1.9.0-3 Manage MD devices aka Linux Software Raid I hope this helps! Let me know if you'd like any more information. Any thoughts on what this might be? Regards, Paul