http://bugzilla.opensuse.org/show_bug.cgi?id=964969 http://bugzilla.opensuse.org/show_bug.cgi?id=964969#c1 David Sterba changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dsterba@suse.com, | |Ulrich.Windl@rz.uni-regensb | |urg.de Flags| |needinfo?(Ulrich.Windl@rz.u | |ni-regensburg.de) --- Comment #1 from David Sterba --- The maximum number of inodes per subvolume is roughly 2^64, I don't think you've run out of them. The numbers reported by df -i is always 0 for inodes/iused/ifree. This is a valid value for filesystems that cannot determine the exact number or do not export such statistics. If you got out of space, that's probably because the metadata space is exhausted. Please paste output of * btrfs fi df /mnt * df -h /mnt -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=964969 http://bugzilla.opensuse.org/show_bug.cgi?id=964969#c2 Ulrich Windl changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(Ulrich.Windl@rz.u | |ni-regensburg.de) | --- Comment #2 from Ulrich Windl --- Explanation first: What I have is a standard btrfs root layout, so the filesystem in question may be /. However / wasn't written, when the problem occurred, but /var/lib/dp/big (another btrfs filesystem without subvolumes). For completeness I give you the stats for all btrfs filesystems. Note what happens with /test/raid1! s1:~ # btrfs fi df / Data, single: total=13.94GiB, used=8.79GiB System, DUP: total=32.00MiB, used=16.00KiB Metadata, DUP: total=1.00GiB, used=799.77MiB GlobalReserve, single: total=224.00MiB, used=0.00B So this looks fine, right? s1:~ # btrfs fi df /var/lib/dp/big Data, single: total=14.00GiB, used=13.72GiB System, DUP: total=8.00MiB, used=16.00KiB System, single: total=4.00MiB, used=0.00B Metadata, DUP: total=1.00GiB, used=24.09MiB Metadata, single: total=8.00MiB, used=0.00B GlobalReserve, single: total=16.00MiB, used=0.00B s1:~ # btrfs fi df /var/lib/dp/lib Data, single: total=5.01GiB, used=2.25GiB System, DUP: total=8.00MiB, used=16.00KiB System, single: total=4.00MiB, used=0.00B Metadata, DUP: total=1.00GiB, used=3.31MiB Metadata, single: total=8.00MiB, used=0.00B GlobalReserve, single: total=16.00MiB, used=0.00B s1:~ # btrfs fi df /var/lib/dp/lib2 Data, single: total=3.01GiB, used=2.26GiB System, DUP: total=8.00MiB, used=16.00KiB System, single: total=4.00MiB, used=0.00B Metadata, DUP: total=1.00GiB, used=2.88MiB Metadata, single: total=8.00MiB, used=0.00B GlobalReserve, single: total=16.00MiB, used=0.00B s1:~ # btrfs fi df /test/raid1 ERROR: couldn't get space info - Inappropriate ioctl for device ERROR: get_df failed Inappropriate ioctl for device s1:~ # btrfs fi df /test/raid2 Data, single: total=5.01GiB, used=4.41GiB System, DUP: total=8.00MiB, used=16.00KiB System, single: total=4.00MiB, used=0.00B Metadata, DUP: total=1.00GiB, used=7.25MiB Metadata, single: total=8.00MiB, used=0.00B GlobalReserve, single: total=16.00MiB, used=0.00B s1:~ # btrfs fi df /test/raid3 Data, single: total=5.01GiB, used=4.41GiB System, DUP: total=8.00MiB, used=16.00KiB System, single: total=4.00MiB, used=0.00B Metadata, DUP: total=1.00GiB, used=7.25MiB Metadata, single: total=8.00MiB, used=0.00B GlobalReserve, single: total=16.00MiB, used=0.00B s1:~ # mount /test/raid1 mount: /dev/mapper/FC70R10-lvol1 is already mounted or /test/raid1 busy /dev/mapper/FC70R10-lvol1 is already mounted on /test/raid1 s1:~ # btrfs fi df /test/raid1 ERROR: couldn't get space info - Inappropriate ioctl for device ERROR: get_df failed Inappropriate ioctl for device s1:~ # umount /test/raid1 s1:~ # mount /test/raid1 s1:~ # btrfs fi df /test/raid1 ERROR: couldn't get space info - Inappropriate ioctl for device ERROR: get_df failed Inappropriate ioctl for device #related dmesg output: [75548.420306] XFS (dm-5): Unmounting Filesystem [75551.146007] XFS (dm-5): Mounting V5 Filesystem [75551.290050] XFS (dm-5): Ending clean mount s1:~ # df -h / /var/lib/dp/* /test/raid* Filesystem Size Used Avail Use% Mounted on /dev/mapper/sys-root 16G 11G 5.2G 68% / /dev/mapper/play-big 266G 14G 250G 6% /var/lib/dp/big /dev/mapper/play-lib 100G 2.3G 96G 3% /var/lib/dp/lib /dev/mapper/play-lib2 48G 2.3G 44G 5% /var/lib/dp/lib2 /dev/mapper/FC70R10-lvol1 274G 4.7G 269G 2% /test/raid1 /dev/mapper/FC300R10-lvol1 3.2T 4.5G 3.2T 1% /test/raid2 /dev/mapper/FC300FR10-lvol1 3.3T 4.5G 3.3T 1% /test/raid3 -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=964969 http://bugzilla.opensuse.org/show_bug.cgi?id=964969#c4 --- Comment #4 from Ulrich Windl --- When mounting another BTRFS filesystem in the same rescue system, the message about "btrfs_uuid_scan_kthread failed -28" does not appear, and I can write to that filesystem. So my guess is: Something on the disk image of the BTRFS prevents a clean mount, and "btrfs check" does not detect that. Could it be worse? -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=964969 Ulrich Windl changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|btrfs ran out of i-nodes, |btrfs ran out of i-nodes (I |but no kernel message |guess), but no kernel |produced |message produced -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.opensuse.org/show_bug.cgi?id=964969 http://bugzilla.opensuse.org/show_bug.cgi?id=964969#c7 --- Comment #7 from Ulrich Windl --- When trying a "btrfs rescue super-recover ..." from the rescue system, I get: Superblock bytenr is larger than device size Usage or syntax errors Then I see a segfault of btrfs use process. Likewise "chunk-recover" produces: read superblock error recover prepare error Fail to recover the chunk tree. Eventually I tries a "zero-log" on the device which succeeded. Unfortunately even after that I get the same message as in comment 4, and any attempt for "filesystem balance" just gives "No space left on device". Next thing I tried was "filesystem defragment" which succeeded, but didn't free up any space for balance. As I had some space left on the VG, I resized the 16GB lvol to 18 GB, and then I could rebalance! To summarize: BTRFS and it tools have to go back to the drawing board for a re-design! -- You are receiving this mail because: You are on the CC list for the bug.