https://bugzilla.novell.com/show_bug.cgi?id=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c1 Jiri Srain changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO InfoProvider| |rriemann@physik.hu-berlin.d | |e --- Comment #1 from Jiri Srain 2011-11-01 07:40:54 UTC --- Did you create a separate /boot partition? Otherwise, it is kind-of expected behavior that your system does not boot... -- 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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c3 Jiri Srain changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO InfoProvider| |rriemann@physik.hu-berlin.d | |e --- Comment #3 from Jiri Srain 2011-11-01 15:00:36 UTC --- OK. Could you attach the YaST installation log (they are stored in the system in /var/log/YaST2) if you can mount the filesystem and get them? -- 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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c6 Arvin Schnell changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW CC| |aschnell@suse.com InfoProvider|rriemann@physik.hu-berlin.d | |e | AssignedTo|aschnell@suse.com |jsrain@suse.com --- Comment #6 from Arvin Schnell 2011-11-02 10:23:22 UTC --- I cannot see why this should be a problem of the YaST partitioner. The Ubuntu forum indicates a missing crc32 module in the initrd. -- 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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c Jiri Srain changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|jsrain@suse.com |mmarek@suse.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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c9 Michal Marek changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dsterba@suse.com --- Comment #9 from Michal Marek 2011-11-02 14:29:58 UTC --- I think the problem is that there is no special handling for btrfs in mkinitrd. If btrfs can use multiple block devices, it needs similar handling like md, dm, etc. David? -- 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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c10 --- Comment #10 from David Sterba 2011-11-09 18:54:19 CET --- I have updated btrfsprogs in factory which will provide the '/sbin/btrfs' command from inside the initrd if the root filesystem is btrfs. The message from syslog in the report

btrfs: faild to read the system array on sda2 btrfs: open_ctree failed

means that there is a multidevice filesystem but the kernel module is not aware of all the devices (besides the one given to mount). Solution 1: run 'btrfs device scan' before the mount; this will scan all devices formated with btrfs filesystem Solution 2: explicitly name all the devices in mount options like: "mount -o device=/dev/sda1,device=/dev/sdb1,device=...,otheroptions /dev S1 is probably the best option as it's generic and you will not end up with unbootable system after you add a new device but do not add it to the mount option list (as would happen in S2 case). Documentation note: specifying a device just means that the kernel module will scan it and add it to it's internal list of known btrfs devices, but only devices belong to the mounted device (ie. the one holding the superblock and FSID) will be tied togather and mounted. IOW, adding any existing btrfs device will do no harm. -- 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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c12 J Powell changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |waveclaw@hotmail.com Version|RC 1 |Final OS/Version|Other |SuSE Other --- Comment #12 from J Powell 2011-12-03 02:10:33 UTC --- I am probably duplicating the btfsprogs work, but wanted to document a solution for 12.1 Final. I resolved this with my 12.1 install by creating the following script as /lib/mkinitrd/scripts/boot-btrfs.sh and re-running mkinitrd. #!/bin/bash #%stage: block #%programs: /sbin/btrfs # ##### btrfs script ## initramfs script for btrfs ## based on http://ubuntu-ky.ubuntuforums.org/showthread.php?p=8784127 ## ## Copyright 2011 assigned to Free Software Foundation ## Licensed under the GNU General Public License 2 or later. ## ## This is where btrfs multi-device filesystems get detected ## ## Command line parameters ## ----------------------- ## ## modprobe btrfs if [ -x /sbin/btrfs ]; then /sbin/btrfs device scan 2>/dev/null fi ## END Note that I can use this to create a new install that has a multi-device btrfs /. Steps to Reproduce a working multidevice btrfs installation of openSuSE 12.1. 1. Boot install CD to rescue mode. 2. Pre-create a root that is btrfs RAID10 (with subvolumes if desired) and a ext4 /boot partition 3. Perform a normal installation. Do not format the btrfs volume, but set the mountpoint to / (or where ever you need) 4. On reboot, boot to install CD to rescue mode 5. Do 'btrfs device scan' to pickup the multi-device btrfs file system 6. Mount root on a convenient location 7. Mount /boot inside that location 8. Copy essential /dev and /proc files from the rescue environment into location 7. chroot to the location / is mounted on 8. Add the script to /lib/mkinitrd/scripts/ 9. mkinitrd to build and re-install a fresh Initrd to the /boot 10. Exit chroot 11. umount all mounted file systems 12. Reboot to continue with Installation I have done this 3 times now on various hardware. Latest 12.1 installation is on a RAID10 / with subvolumes using 4x 1Tb SATA drives from which I am posting this. Installer does not appear to do a proper RAID setup beyond JBOD/RAID0 or give the user options for that. Installer can use an existing RAID manually setup. Of course, older systems would need to use btrfsctl and issue 'btrfsctl -a' as that deprecated mechanism does the device scan. -- 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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c Michal Marek changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW InfoProvider|dsterba@suse.com | AssignedTo|mmarek@suse.com |dsterba@suse.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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c15 David Sterba changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED --- Comment #15 from David Sterba 2012-02-08 16:27:13 CET --- Something is wrong with the boot 'dev scan' step. I installed fresh SP2 with root btrfs on one device. Then added a few more disks to the root device and rebooted. Boot failed, the other devices were not detected due to missing 'dev scan' step at that time. Stuck inside the initrd, I was not able to fix it by hand other than by reducing size of the device group of / back to one device (time consuming due to needed data shuffling). I tried to manually add all devices as a mount option, and then reboot. Did not work either, contents of /etc/fstab in the initrd environment did not have the same mount options. /lib/mkinitrd/boot # ls -1 01-acpi.sh 01-devfunctions.sh 01-fv_guest.sh 01-vendor.sh 02-start.sh 03-dm.sh 03-scsi_dh.sh 03-storage.sh 04-multipathd.sh 04-udev.sh 05-blogd.sh 05-clock.sh 05-kms.sh 11-block.sh 11-firewire.sh 11-ibft.sh 11-usb.sh 12-network.sh 13-netconsole.sh 21-btrfs.sh <--- do 'dev scan' here 21-cifs.sh 21-devinit_done.sh 21-dmraid.sh <--- what if i have btrfs on top of DM ? 21-multipath.sh 21-nfs.sh 22-kpartx.sh 51-md.sh <--- or MD? 61-lvm2.sh 71-luks.sh <--- or I'm living dangerously on an encrypted volume? 81-resume.userspace.sh 82-resume.kernel.sh 83-mount.sh 84-remount.sh 91-createfb.sh 91-kdump.sh 91-killblogd.sh 91-killudev.sh 91-shell.sh 92-killblogd2.sh 92-killmultipathd.sh 93-boot.sh --- 21-btrfs.sh has #%stage: block and no dependencies, but I think it should be run either just before '83-mount' or '81-resume'. Is there any dependency target meaning 'after all block devs are up AND before mount' ? -- 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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c18 --- Comment #18 from David Sterba 2012-03-07 18:10:03 CET --- I have updated factory btrfsprogs and submitted update to 12.1. The initrd boot scripts should be run just before the mount step. mkinitrd has to be rerun so the initrd's are rebuilt with this change. -- 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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c20 Swamp Workflow Management changed: What |Removed |Added ---------------------------------------------------------------------------- Status Whiteboard|maint:running:46605:moderat |maint:running:46605:moderat |e |e | |maint:released:sle11-sp2:46 | |825 --- Comment #20 from Swamp Workflow Management 2012-05-30 17:03:44 UTC --- Update released for: btrfsprogs, btrfsprogs-debuginfo, btrfsprogs-debugsource Products: SLE-DEBUGINFO 11-SP2 (i386, ia64, ppc64, s390x, x86_64) SLE-DESKTOP 11-SP2 (i386, x86_64) SLE-SERVER 11-SP2 (i386, ia64, ppc64, s390x, x86_64) SLES4VMWARE 11-SP2 (i386, x86_64) -- 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=727383 https://bugzilla.novell.com/show_bug.cgi?id=727383#c21 David Sterba changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #21 from David Sterba 2013-11-25 17:54:05 CET --- Closing. -- 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.