[Bug 656217] New: Software RAID Devices on AHCI Drives Fail On Reboot
https://bugzilla.novell.com/show_bug.cgi?id=656217 https://bugzilla.novell.com/show_bug.cgi?id=656217#c0 Summary: Software RAID Devices on AHCI Drives Fail On Reboot Classification: openSUSE Product: openSUSE 11.3 Version: Final Platform: PC OS/Version: Other Status: NEW Severity: Major Priority: P5 - None Component: Basesystem AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: malv_star@hotmail.com QAContact: qa@suse.de Found By: --- Blocker: --- Created an attachment (id=402392) --> (http://bugzilla.novell.com/attachment.cgi?id=402392) Dmesg output for failed boot. User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-GB; rv:1.9.2.12) Gecko/20101026 SUSE/3.6.12-0.7.1 Firefox/3.6.12 Creating a software RAID during install, on drives that are set in BIOS to use AHCI, causes these RAID devices to become inaccessible after first reboot. For example, creating a mirrored RAID during install (say /dev/md0) and mounting it at, for example /home, will causes the "repair filesystem" command-line prompt to display during boot. The system claims that fsck failed for one or more devices, and that the filesystem must be repaired manually. In fact, the device is perfectly accessible when references as /dev/md0, but using /dev/disk/by-id/md-xxxxx it is not. Changing the fstab entry from using /dev/disk/by-id to /dev/md0 allowed the system to boot perfectly normally. Further details follow, plus attachments. This problem was observed with multiple file systems on the RAID device, on multiple machines. Reproducible: Always Steps to Reproduce: 1. Create a mirrored software RAID during install 2. Proceed normally, and after between one and three reboots, the RAID will become inaccessible on boot. Actual Results: The software RAID is no longer mounted at boot time. Expected Results: The software RAID should be mounted normally. A workaround is to alter the fstab to set the system to search for /dev/md0 instead of using the /dev/disk/by-id reference. -- 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=656217 https://bugzilla.novell.com/show_bug.cgi?id=656217#c1 --- Comment #1 from Malvern Star <malv_star@hotmail.com> 2010-11-29 07:38:15 UTC --- Created an attachment (id=402393) --> (http://bugzilla.novell.com/attachment.cgi?id=402393) Directory listing for /dev/disk/by-id/ folder -- 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=656217 https://bugzilla.novell.com/show_bug.cgi?id=656217#c2 --- Comment #2 from Malvern Star <malv_star@hotmail.com> 2010-11-29 07:39:16 UTC --- Created an attachment (id=402394) --> (http://bugzilla.novell.com/attachment.cgi?id=402394) File system table for unbootable system. Altering the final line of the attached fstab to use /dev/md0 functions as a workaround. The system boots normally. -- 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=656217 https://bugzilla.novell.com/show_bug.cgi?id=656217#c Xinli Niu <xlniu@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED CC| |xlniu@novell.com AssignedTo|bnc-team-screening@forge.pr |nfbrown@novell.com |ovo.novell.com | -- 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=656217 https://bugzilla.novell.com/show_bug.cgi?id=656217#c3 Neil Brown <nfbrown@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO InfoProvider| |malv_star@hotmail.com --- Comment #3 from Neil Brown <nfbrown@novell.com> 2010-11-29 09:56:01 UTC --- Thanks for the report. Based on your description there seems to be some sort of race with device names appearing in /dev, particularly as it doesn't always fail, but does often. The script /etc/init.d/boot.localfs is responsible for getting this right. It is the script that causes the "repair filesystem" prompt to appear. That script has a function 'wait_for_udev' which should wait for the link to appear. You should see a message like: Waiting for /dev/disk/by-id/md-uuid-d87fe1ec:5ddde35d:eab631e5:c6c4a0ad ... After 30 '.'s it should should say 'timeout!', otherwise it should succeed. Do you see anything like that during boot. These messages don't appear in any logs unfortunately, only on the console. A digital photo of the screen when it stops and gives you a prompt might be quite useful. -- 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=656217 https://bugzilla.novell.com/show_bug.cgi?id=656217#c4 --- Comment #4 from Malvern Star <malv_star@hotmail.com> 2010-11-29 14:53:11 UTC --- I can't seem to give that to you. Now that I have used /dev/md0, changing back the fstab to use /dev/disk/by-id seems to be completely ignored - as far as I can tell, the system is continuing to mount using /dev/md0. I tried installing another system, using /dev/md0 for the root partition and GRUB gave Error 17: Cannot mount selected partition on the first reboot. Sure enough, it's looking for root=/dev/disk/by-id I'll try another approach, but you may end up having to attempt this yourself. -- 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=656217 https://bugzilla.novell.com/show_bug.cgi?id=656217#c5 Malvern Star <malv_star@hotmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED InfoProvider|malv_star@hotmail.com | --- Comment #5 from Malvern Star <malv_star@hotmail.com> 2010-12-13 17:43:59 UTC --- OK, I can't reproduce this in 11.4 M4, but I have some further information for 11.3 Final. I don't think the problem is a race condition, because when the RAID holds /, GRUB reports that it is unable to find the root partition using /dev/disk/by-id. So we don't even get that far. The problem with recognising MD devices on AHCI seems to be much more fundamental. -- 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=656217 https://bugzilla.novell.com/show_bug.cgi?id=656217#c6 Neil Brown <nfbrown@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |WORKSFORME --- Comment #6 from Neil Brown <nfbrown@novell.com> 2011-03-24 00:25:06 UTC --- As you cannot reproduce this in 11.4, I think I'll just close it. I'm not sure I would expect GRUB to understand /dev/disk/my-id/.... names, but I don't know the ins-and-outs of grub very well so I could be wrong. Thanks for the report anyway - sorry we couldn't get to the bottom of it. -- 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=656217 https://bugzilla.novell.com/show_bug.cgi?id=656217#c7 --- Comment #7 from Malvern Star <malv_star@hotmail.com> 2011-04-03 04:52:05 UTC --- I can confirm your findings for 11.4 final. It has been suggested that 11.3 has "device enumeration issues", but beyond that I can't say much more. I'm inclined to agree with closing this bug as it is hopefully no longer a problem. Thanks, Solaris. -- 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.
participants (1)
-
bugzilla_noreply＠novell.com