Bug 111161 - bogus 'some devices missing' prevents boot
Summary: bogus 'some devices missing' prevents boot
Status: NEW
Alias: None
Product: File System
Classification: Unclassified
Component: btrfs (show other bugs)
Hardware: All Linux
: P1 high
Assignee: Josef Bacik
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-01-22 03:38 UTC by Konstantin Svist
Modified: 2016-03-20 09:57 UTC (History)
1 user (show)

See Also:
Kernel Version: 4.2.3-300.fc23.x86_64
Subsystem:
Regression: No
Bisected commit-id:

Attachments
copy of bad array -> rm all files/subvolumes -> btrfs-image -c9 (5.00 KB, application/octet-stream)
2016-01-23 19:06 UTC, Konstantin Svist 		Details
output of debug-tree -d (7.80 KB, text/plain)
2016-01-24 02:11 UTC, Konstantin Svist 		Details
output of debug-tree -e (4.79 KB, text/plain)
2016-01-24 02:11 UTC, Konstantin Svist 		Details
Add an attachment (proposed patch, testcase, etc.)

Description Konstantin Svist 2016-01-22 03:38:25 UTC 
- mounts OK without degraded
- btrfs check does not report missing devices
- btrfs scrub completed without problems, no change
- btrfs fi show does not report missing devices
- btrfs fi show -d does report missing devices
- btrfs dev del missing reports there are no missing devices
- btrfs dev ready exit value 1, so boot always fails (systemd waits without timeout until exit value is 0)

Boot kernel version: 4.3.3-300.fc23.x86_64
Mounting okay from Fedora 23 x64 LXDE Live (kernel 4.2.3-300.fc23.x86_64)
btrfs-progs v4.3.1

Problem appeared after replacing a bad drive (btrfs dev add, then btrfs dev del)
All drives are okay right now & btrfs is balanced.

# btrfs fi df /mnt/
Data, RAID10: total=1.29TiB, used=1.29TiB
System, RAID10: total=64.00MiB, used=160.00KiB
Metadata, RAID10: total=4.00GiB, used=3.12GiB
GlobalReserve, single: total=512.00MiB, used=0.00B


# btrfs fi show /mnt/
Label: none  uuid: 48f0e952-a176-481e-a184-6ee51acf54b1
        Total devices 4 FS bytes used 1.29TiB
        devid    1 size 931.51GiB used 664.03GiB path /dev/sda2
        devid    3 size 931.51GiB used 664.03GiB path /dev/sdc2
        devid    4 size 931.51GiB used 664.03GiB path /dev/sdd2
        devid    6 size 931.51GiB used 664.03GiB path /dev/sdb2

# btrfs fi show -d
Label: none  uuid: 48f0e952-a176-481e-a184-6ee51acf54b1
        Total devices 5 FS bytes used 1.29TiB
        devid    1 size 931.51GiB used 664.03GiB path /dev/sda2
        devid    3 size 931.51GiB used 664.03GiB path /dev/sdc2
        devid    4 size 931.51GiB used 664.03GiB path /dev/sdd2
        devid    6 size 931.51GiB used 664.03GiB path /dev/sdb2
        *** Some devices missing

dmesg:
http://cache.gmane.org//gmane/comp/file-systems/btrfs/52437-002.bin

http://thread.gmane.org/gmane.comp.file-systems.btrfs/52437
Comment 1 Konstantin Svist 2016-01-23 19:06:21 UTC 
Created attachment 201121 [details]
copy of bad array -> rm all files/subvolumes -> btrfs-image -c9

I've made a dd copy of the problematic array, changed the UUID in the copy and whittled it down to 0 files (removed all subvolumes)

As per advice in #btrfs, here's an image of said 0-file array, made by "btrfs-image -c9"

Please let me know ASAP if this is sufficient to see the error or if I need to send in more info.

I'd like to restore my main array ASAP
Comment 2 Konstantin Svist 2016-01-23 19:07:36 UTC 
Note: just in case, the new UUID is 367419ac-01df-4432-a5fb-1483c49abc46
Comment 3 Konstantin Svist 2016-01-24 02:11:34 UTC 
Created attachment 201191 [details]
output of debug-tree -d
Comment 4 Konstantin Svist 2016-01-24 02:11:56 UTC 
Created attachment 201201 [details]
output of debug-tree -e
Note You need to log in before you can comment on or make changes to this bug.