Mkfs tends to create pretty large metadata chunk compared to kernel:
Node size: 16384
Sector size: 4096
Filesystem size: 10.00GiB
Block group profiles:
Data: single 8.00MiB
Metadata: DUP 1.00GiB
System: DUP 8.00MiB
While kernel only tends to create 256MiB metadata chunk:
/* for larger filesystems, use larger metadata chunks */
if (fs_devices->total_rw_bytes > 50ULL * SZ_1G)
max_stripe_size = SZ_1G;
else
max_stripe_size = SZ_256M;
This won't cause problems in real world, but it's still better to make
the behavior unified.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Unlike kernel, btrfs-progs doesn't (yet) support devices grow/shrink,
the port only needs to handle open_ctree() time initialization (at
read_one_dev()), and btrfs_add_device() used for mkfs.
This provide the basis for incoming unification of chunk allocator
behavior.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This patch adds option --forget to 'device scan'
$ btrfs device scan --forget [dev...]
to unregister the given device from kernel module. The device cannot be
part of a mounted filesystem, this will be reported as an error.
If no argument is given it will unregister all stale (device which are
not mounted) from the kernel.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Anand Jain <anand.jain@oracle.com>
[ update changelog ]
Signed-off-by: David Sterba <dsterba@suse.com>
This reverts commit cb8abddb20.
There are several reports in IRC that this patch breaks in some
send/receive environments. There are no exact steps to reproduce, only
approximate descroptions. Until a proper reproducer is known, the patch
is temporarily reverted due to the user-visible impact.
Issue: #162
Signed-off-by: David Sterba <dsterba@suse.com>
Since this row is longer than the rest it's enough to use a single tab
character to delimig the value from caption. Without this patch
output looked like:
nodesize 16384
leafsize (deprecated) 16384
With it:
nodesize 16384
leafsize (deprecated) 16384
No functional changes just making the output a bit neater.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Use the Xenial dist for CI so now we have only 2 years old base version,
instead of 4 years old. Kernel is 4.15, gcc is 5.4, clang 7.0.
Signed-off-by: David Sterba <dsterba@suse.com>
The autodetection of conversion filesystems support will skip the tests
if the library is not available for some reason. This now happens for
reiserfs in the CI environment. We need to know about such case so
request the support.
Signed-off-by: David Sterba <dsterba@suse.com>
For some reason the test was not enabled in the CI builds and would have
caught the build problem fixed by the previous patch.
Signed-off-by: David Sterba <dsterba@suse.com>
Commit 75b5eabb61 ("btrfs-progs: uuid: Port
btrfs_uuid_tree_add() function") brings code from kernel and introduces
an unsatisfied build dependency for libbtrfs. This fails build for
external library users like snapper.
As a hotfix, shuffle the new UUID functions so the build works. Simply
adding uuid-tree.o to libbtrfs_objects does not work due to other
missing symbols.
[CC] uuid-tree.o
[LD] libbtrfs.so.0.1
[LN] libbtrfs.so
[TEST PREP] library-test
ld: .../libbtrfs.so: undefined reference to `btrfs_alloc_path'
ld: .../libbtrfs.so: undefined reference to `btrfs_extend_item'
ld: .../libbtrfs.so: undefined reference to `btrfs_insert_empty_items'
ld: .../libbtrfs.so: undefined reference to `btrfs_free_path'
ld: .../libbtrfs.so: undefined reference to `read_extent_buffer'
ld: .../libbtrfs.so: undefined reference to `btrfs_mark_buffer_dirty'
ld: .../libbtrfs.so: undefined reference to `write_extent_buffer'
ld: .../libbtrfs.so: undefined reference to `btrfs_search_slot'
LinK: https://bugs.gentoo.org/675974
Link: https://github.com/openSUSE/snapper/issues/473
Fixes: 75b5eabb61 ("btrfs-progs: uuid: Port btrfs_uuid_tree_add() function")
Signed-off-by: David Sterba <dsterba@suse.com>
Reformat and update the help string so it's more aligned with other
utilities (mkfs.btrfs was taken as a source):
- reduce indentation
- group related options
- add references to mkfs features
- wording adjustments
Signed-off-by: David Sterba <dsterba@suse.com>
Enhance the note what scrub does not compared to 'check'.
Author: Andrea Gelmini <andrea.gelmini@gelma.net>
Signed-off-by: David Sterba <dsterba@suse.com>
Commit e578b59bf6 ("btrfs-progs: convert strerror to implicit %m")
missed adding braces after a conditional so we will see the following
message whenever a tree block needs repair, regardless of whether repair
was successful: "Failed to repair btree: Success"
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Starting device replace and balance does not properly report the status.
1. start replace:
$ btrfs replace start -B -f /dev/sdd /dev/sdb /btrfs
2. while replace is still running, from another terminal try to start
balance:
$ btrfs balance start --full-balance /btrfs
Done, had to relocate 0 out of 0 chunks
This returns with exit code 0, which fails to report that balance failed
to start because another exclusive operation is running.
In fact kernel ioctl BTRFS_IOC_BALANCE(_V2) does return error code 8,
but it's incorrectly reset to 0. Fix it by checking for the error code > 0.
After:
$ btrfs balance start --full-balance /btrfs
ERROR: balance: add/delete/balance/replace/resize operation in progress
Signed-off-by: Anand Jain <anand.jain@oracle.com>
[ update changelog ]
Signed-off-by: David Sterba <dsterba@suse.com>
Add a test for a scenario that used to fail due to find_mount_root()
incorrectly determining the mount point for the receive path due to the
fact that a different mount point with a path that is a prefix of the
receive path exists.
This is fixed by a recent patch titled:
"Btrfs-progs: fix mount point detection due to partial prefix match"
Reported-by: David Disseldorp <ddiss@suse.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
[ adjustments to make the test work when hosted on NFS ]
Signed-off-by: David Sterba <dsterba@suse.com>
The strdup() function can fail, in which case it returns NULL and sets
errno [1]. Therefore add the missing error check after calling strdup()
at find_mount_root().
[1] - http://man7.org/linux/man-pages/man3/strdup.3p.html
Reviewed-by: David Disseldorp <ddiss@suse.de>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When attempting to find the mount point of a path we can end up returning
an incorrect mount point. This happens because we consider a mount point
valid for the given path even if it only partially matches the path.
Consider the following example, which makes btrfs receive fail:
$ truncate -s 1G disk1
$ truncate -s 1G disk2
$ losetup /dev/loop1 disk1
$ losetup /dev/loop2 disk2
$ mkfs.btrfs -f /dev/loop1
$ mkfs.btrfs -f /dev/loop2
$ mount /dev/loop1 /mnt
$ mkdir /mnt/ddis
$ mkdir /mnt/ddis-not-a-mount
$ mount /dev/loop2 /mnt/ddis
$ echo "some data" > /mnt/ddis/file
$ btrfs subvolume snapshot -r /mnt/ddis /mnt/ddis/snap
$ btrfs send -f /tmp/send.data /mnt/ddis/snap
$ btrfs receive -f /tmp/send.data /mnt/ddis-not-a-mount
At subvol snap
ERROR: chown failed: No such file or directory
In that example btrfs receive passes the path "/mnt/ddis-not-a-mount" to
find_mount_root() which picks "/mnt/ddis" as the mount point instead of
"/mnt". The wrong decision happens because "/mnt/ddis" is the longest
string found that is a prefix of "/mnt/ddis-not-a-mount", however it
shouldn't be considered valid because what follows the substring "ddis"
in the given path is not a path separator ("/") nor the null character
('\0'). So fix find_mount_root() to check for the presence of a path
separator or a null byte character after if finds a mount point string
that matches the given path.
A test case will follow soon in a separate patch.
Reported-by: David Disseldorp <ddiss@suse.com>
Reviewed-by: David Disseldorp <ddiss@suse.de>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Provide an 'etags' make target to create tags in the Emacs etags
format, similar to the 'tags' target for VIM's ctags.
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>
This function provides the offline functionality to add new uuid tree
entry. Also port fs_info->uuid and its initialization and cleanup code
to support uuid tree.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Although we have btrfs_uuid_tree_lookup_any(), it's an online function
utilizing tree search ioctl, not an offline search function.
This patch will port kernel btrfs_uuid_tree_lookup() into btrfs-progs
for later proper uuid tree initialization.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For data reloc tree creation, we copy its contents from the fs tree just
for its INODE_ITEM, INODE_REF and dirid. This hides the detail and is
not obvious for why we're copying from fs root.
This patch will create data reloc tree from scratch:
- Create root, including root item and new tree root
- Change dirid to BTRFS_FIRST_FREE_OBJECTID
- Insert root INODE_ITEM and INODE_REF
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Just as how kernel uses it.
This provides the basis for later uuid creation.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Add the section with brief summary of the support. The supported
features could be mentioned explicitly with versions, and other examples
of "check before use" that could lead to unbootable systems.
Signed-off-by: David Sterba <dsterba@suse.com>
The runtime of the test is over 4 minutes when hosted on NFS, the
fallocate phase takes the most time. As fallocate can't take multiple
arguments, we can't save the overhead by creating the files at once.
Create the files in batches in the background, this improves the runtime
to 2 minutes.
Signed-off-by: David Sterba <dsterba@suse.com>
The repair function is reusing delete_corrupted_dir_item().
Since the error can happen for root dir inode, also call
try_repair_inode() on root dir inode.
This is especially important for old filesystems, since later kernel
introduces stricter tree-checker, which could detect such hash mismatch
and refuse to read the corrupted leaf.
With this repair ability, user could repair with btrfs check --repair.
Link: https://bugzilla.opensuse.org/show_bug.cgi?id=1111991
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This changes reporting from current in-place, like:
ERROR: DIR_ITEM[256 751495445] name foor.WvG1c1TdU namelen 13 filetype 1 mismatch with its hash, wanted 751495445 have 2870353892
root 5 root dir 256 error
To new summary report at the end of the pass:
root 5 root dir 256 error
root 5 inode 256 errors 40000
Dir items with mismatch hash:
name: foor.WvG1c1Td namelen: 13 wanted 0xab161fe4 has 0x2ccae915
Also, with mismatch_dir_hash_record structure, it provides the base for
later original mode repair.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For DIR_ITEM with mismatch hash, we could just remove the offending dir
item from the tree.
Lowmem mode will handle the rest, either re-create the correct dir_item
or move the orphan inode to lost+found.
This is especially important for old filesystems, since later kernel
introduces stricter tree-checker, which could detect such hash mismatch
and refuse to read the corrupted leaf.
With this repair ability, user could repair with 'btrfs check
--mode=lowmem --repair'.
Link: https://bugzilla.opensuse.org/show_bug.cgi?id=1111991
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There's a suggestion in #158 to improve the documenatation of the subvol
show command so it's more obvious what kind of information it provides.
I've updated the docs according to that and added an example that should
make the request for creation time at least greppable, the overall
description mentions 'times' so this should be enough as a pointer.
Pull-request: #158
Suggested-by: Robert Pollak <robert.pollak@posteo.net>
Signed-off-by: David Sterba <dsterba@suse.com>
Now two locations can detect such problem, either by device item
used/total bytes check, or by early dev extents check against device
boundary.
The image is hand-crafted image which uses DATA SINGLE chunk to feed
btrfs check. As expected, as long as block group item, chunk item,
device used bytes match, older versions check can't detect such problem.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
