[Bug 763314] New: RAID 10 created using mdadm not recognized at boot
https://bugzilla.novell.com/show_bug.cgi?id=763314 https://bugzilla.novell.com/show_bug.cgi?id=763314#c0 Summary: RAID 10 created using mdadm not recognized at boot Classification: openSUSE Product: openSUSE 12.2 Version: Milestone 3 Platform: x86-64 OS/Version: openSUSE 12.2 Status: NEW Severity: Normal Priority: P5 - None Component: Bootloader AssignedTo: jsrain@suse.com ReportedBy: robert.munteanu@gmail.com QAContact: jsrain@suse.com Found By: --- Blocker: --- User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:11.0) Gecko/20100101 Firefox/11.0 I have performed a fresh install of 12.2 M3 , with the following setup: / -> /dev/sda1 , ext4 /mnt/md -> /dev/md0 RAID 10 array, ext4 At boot the RAID array is not recognized ; I see some error messages when udev tries to create the array udevd [170]: failed to execute /sbin/mdadm: '/sbin/mdadm --incremental /dev/sdb1 ' : No such file or directory and it's repeated four times - once for each partition in the raid array. Also the boot is delayed for a long time [ 12.572749] md/raid10:md0: active with 4 out of 4 devices [ 12.572777] md0: detected capacity change from 0 to 620806275072 [ 12.824973] md0: unknown partition table [ 97.466489] vboxdrv: Found 4 processor cores. Mounting the partition after the boot sequence is complete works like a charm. The configured bootloader is grub2 . This setup worked fine on 12.1 . Reproducible: Always -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=763314
https://bugzilla.novell.com/show_bug.cgi?id=763314#c
Jiri Srain
https://bugzilla.novell.com/show_bug.cgi?id=763314
https://bugzilla.novell.com/show_bug.cgi?id=763314#c1
Neil Brown
https://bugzilla.novell.com/show_bug.cgi?id=763314
https://bugzilla.novell.com/show_bug.cgi?id=763314#c2
Robert Munteanu
If you remove the line 64-md-raid.rules \
from /lib/mkinitrd/scripts/setup-dev.sh and run mkinitrd again, then reboot, does that make a difference?
(assuming it's setup-udev.sh ) Yes, that does make a difference. The error messages no longer appear and I can remove the 'noauto' option from /etc/fstab . Previously the array would be partially assembled ( 3/4 disks ) and would stop booting. I failed to mention that, but that bothered me the most, not the error messages. It's interesting to see that in the dmesg output the first three partitions are bound way ahead of the fourth [ 11.948857] md: bind<sdc1> [ 11.951483] md: bind<sdb1> [ 11.959807] md: bind<sdd1> .... [ 13.538173] md: bind<sda2> [ 13.564037] md: raid10 personality registered for level 10 Anyway, for the immediate term this is fixed.
As your md array is not required for boot, the initrd should not try to assemble it, and because it doesn't need to assemble it, it doesn't contain mdadm. However it does containt the udev script, hence the error messages.
So the error messages are not good, but not indicative of a serious problem.
The 85 second delay is more worrying. However there is no evidence that that is caused by md. Can you try booting with "debug=yes" on the kernel command line? I think it is fairly easy to edit the command line with grub2. That should print on the console which script is being executed so you should be able to see better where the delay is.
Booting with debug=true ( in the grub2 linux ..... line ) still does not give me any clue. The gap is still [ 16.723360] EXT4-fs (md0): mounted filesystem with ordered data mode. Opts: (null) [ 96.562483] vboxdrv: Found 4 processor cores. Unfortunately the scripts are not retained in the dmesg output but IIRC the last print was something like 93-boot.sh . I'll attach the full dmesg log shortly. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=763314
https://bugzilla.novell.com/show_bug.cgi?id=763314#c3
--- Comment #3 from Robert Munteanu
https://bugzilla.novell.com/show_bug.cgi?id=763314
https://bugzilla.novell.com/show_bug.cgi?id=763314#c4
Neil Brown
https://bugzilla.novell.com/show_bug.cgi?id=763314
https://bugzilla.novell.com/show_bug.cgi?id=763314#c5
Neil Brown
https://bugzilla.novell.com/show_bug.cgi?id=763314
https://bugzilla.novell.com/show_bug.cgi?id=763314#c6
--- Comment #6 from Robert Munteanu
https://bugzilla.novell.com/show_bug.cgi?id=763314
https://bugzilla.novell.com/show_bug.cgi?id=763314#c7
--- Comment #7 from Neil Brown
https://bugzilla.novell.com/show_bug.cgi?id=763314
https://bugzilla.novell.com/show_bug.cgi?id=763314#c8
--- Comment #8 from Robert Munteanu
You should be able to grab an RPM from
which should fix the problem. It should turn in Factory and 12.2 in due course.
I've installed the updated RPM ( now at mdadm-3.2.5-61.2.x86_64 ) and with the patched /lib/mkinitrd/scripts/setup-udev.sh booting works fine and the delay is still there. However, once I've corrected the swap entry from /etc/fstab to point to an existing partition ( this was the first line in the file ) the boot delay is gone. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=763314
https://bugzilla.novell.com/show_bug.cgi?id=763314#c9
Neil Brown
participants (1)
-
bugzilla_noreply@novell.com